No. 14 — April 7, 2000

Feature Article


Douglas Ostrom

Issue Index aaaa 2000 Archive Index aaaa Subscriber Area aaaa Home


Whether called keiretsu, zaibatsu or by some other name, corporate groupings have been a distinctive part of Japan's business scene for decades. For almost as long, analysts in the United States and elsewhere have suggested that the days of keiretsu are numbered because the exclusive relationships that such groups imply are thought to be inconsistent with market forces. With Japan's Big Bang financial reform and other liberalization initiatives in progress, this argument has been made more loudly in recent years. For U.S. policymakers as well as for rivals of Japanese firms that consider keiretsu a hindrance to open markets, a keiretsu collapse could not come a day too soon.

Some analysts, particularly those based in Japan, have recognized the economic advantages of keiretsu and, thus, have been reluctant to forecast their demise. Yet they, too, now believe that the role of keiretsu — whether a group clustered around a financial institution or one formed by a manufacturer and its production chain — is diminishing.

Defenders and critics alike have overestimated the economic impact of keiretsu. However, the evidence is somewhat mixed as to whether these corporate groupings are in decline. Linkages among many types of keiretsu firms are weakening for several reasons, but the ties among financial members of keiretsu are becoming stronger. Moreover, even nonfinancial companies are dependent on a smaller number of banks than in the past, which might reflect an increased reliance on keiretsu lenders. Finally, Internet-related businesses talk about forming Net-batsu — a sort of Internet zaibatsu — although the organization of such groups is in its early stages at best.



Keiretsu (corporate groupings) unquestionably are one of the most interesting aspects of Japan's economy. These mammoth organizations seem uniquely equipped to do good or evil on a large scale. Indeed, they regularly have been described as doing both. From the time of the American occupation, if not earlier, keiretsu have held a particular fascination for Americans who believe, quite correctly, that nothing comparable exists on this side of the Pacific.

Keiretsu come in many varieties, shapes and sizes, but in no case do they constitute legal entities. Thus, determining which firms participate is a somewhat subjective process. On paper at least, members are independent companies. No one person or single firm occupies a leadership position in a formal sense. The various Mitsubishi Group companies, for example, are not subsidiaries of some third company or of each other. Shares of the member firms are widely traded, but there is no such thing as Mitsubishi Group stock. Mitsubishi Corp. merely is the trading company component of the Mitsubishi Group and has no special status within the keiretsu.

Keiretsu can be divided into two basic types: financial (horizontal) and nonfinancial (vertical). The difference between them is the glue that holds the group together as well as the strength of that bond. Members of horizontal keiretsu are linked by the power and the obligations of the large bank that sits at the center of the organization and, to a lesser extent, by the group's trading companies; in some instances, they are tied together by a shared history. The cohesiveness of vertical keiretsu, by contrast, depends on one or more large nonfinancial companies that hold substantial equity positions in affiliates and that serve as important customers and suppliers to the rest of the group. Horizontal and vertical keiretsu have evolved differently and have shown various amounts of staying power through the upheavals that have rocked the Japanese economy in recent decades.


Financial Keiretsu

Historical Background - Financial keiretsu like the Mitsubishi Group have as their bonding agent a financial company, usually a large nationwide commercial bank known as a city bank. The Mitsubishi Group's main bank, Bank of Tokyo-Mitsubishi, Ltd., is one of nine such institutions today. The others are Dai-Ichi Kangyo Bank, Ltd., Sanwa Bank, Ltd., Fuji Bank, Ltd., Sakura Bank, Ltd., Tokai Bank, Ltd., Asahi Bank, Ltd., Daiwa Bank, Ltd. and Sumitomo Bank, Ltd. Each of these banks as well as their smaller brethren is the center of a keiretsu, although most analysts focus on the so-called Big Six groups and their banks: Mitsui (which has Sakura Bank at its core), Mitsubishi, Sumitomo, Sanwa, Fuyo (Fuji Bank) and Dai-Ichi.

The first three of these keiretsu have a special role in Japanese history. Mitsui, Mitsubishi and Sumitomo are each direct descendants of a pre-1945 zaibatsu (financial clique). In contrast to modern financial keiretsu, zaibatsu did have central direction, often through a dominant family, exercised via a holding company that had a controlling interest in zaibatsu companies. In that respect, they were similar to the chaebol of today's South Korea.

Gen. Douglas MacArthur and other occupation leaders laid substantial blame on zaibatsu for Japan's militarism in the 1930s and 1940s. During World War II, zaibatsu companies produced a large part of the nation's weaponry — for example, Mitsubishi's A6M Reisen, the infamous Zero fighter plane. However, not a few scholars have concluded that even 60 years ago, zaibatsu represented the establishment in Japan; as such, they clearly saw what they could lose through armed conflict with the United States and its allies. In this view, it was shinzaibatsu (new zaibatsu), anxious to elbow their way into the top ranks of the economic hierarchy, that were more eager for war.

Whether old or new, zaibatsu loomed as large, threatening organizations to occupation authorities, who ordered them dissolved. Thousands of top officials of 245 big concerns were purged, the companies' equity was seized, the holding companies were disbanded and, in a few instances, individual companies were broken up.1

Partly for these reasons, the Japanese economy in the immediate postwar years was characterized by incredible flux and uncertainty. One consequence of that environment was a burst of entrepreneurship to fill the vacuum. Entirely new companies were established that were completely free of zaibatsu ties.2 Sony Corp., founded in 1946, is a prime example.

Holding companies no longer existed to coordinate activities among zaibatsu companies. Moreover, very young, junior-level executives suddenly were in charge of huge firms whose physical facilities in some instances had been severely damaged and whose customers had been impoverished by the war. One study found that of 46 new chief executives whose backgrounds could be determined, only six had experience during the war in positions as high as executive director; 20 never had served on a board of directors. Generally, however, the new presidents had reached their positions through internal promotion and were company veterans despite their youth.3

The young men running corporate Japan turned to one of the few sources of expert help available: their former coworkers from zaibatsu, whom they knew from their rotations through member companies. As early as 1946, executives formalized such contacts through what have become known as Presidents' Clubs. In the meetings of these organizations, which were illegal at the time of their inception, top officials could share such information as what the American occupiers were up to and thrash out problems like how to obtain scarce supplies. The clubs took names like Hakusui-kai (White Water Club) — the name for the Sumitomo Group meetings — that may have reflected their original clandestine natures.

When the American occupation ended in 1952, some zaibatsu reforms were reversed, but others remained in place. For example, the new executives had become comfortable in their positions and never did return control to their former bosses. The Presidents' Clubs took on a more formal existence. Equity in old zaibatsu companies, which, during the occupation, had been widely distributed among employees and others, tended to flow back to zaibatsu partners. Holding companies, however, remained banned until the late 1990s.

The resultant hybrid organizations became known as keiretsu, translated literally as "series" or "linkage." In contrast to the Mitsui, Mitsubishi and Sumitomo keiretsu, the other members of the Big Six — Sanwa, Dai-Ichi and Fuyo — have no direct prewar predecessors. Although their core banks are comparable in terms of size and influence to the trio of zaibatsu descendants, the latter three usually are considered much weaker organizations.

The Big Six Today - Some companies, particularly if they are the result of mergers among members of different keiretsu, belong to two — or even three — Presidents' Clubs (see Table 1). Most analysts argue that more exclusive, smaller groups are more likely to be effective because coordination of members' activities is easier. On that basis, the Mitsubishi Group and the Sumitomo Group would appear to have advantages over other keiretsu, a view roughly matching the consensus.

Today, Presidents' Clubs meet on a regular basis, typically once a month in a downtown Tokyo office, although the Sumitomo Group and the Sanwa Group, both with lead banks headquartered in Osaka, hold some meetings in that city. Presidents only or presidents and chairmen attend, depending on the club. Most meetings are described as vehicles for exchanging views. The Dai-Ichi Group's Presidents' Club, though, just meets quarterly in Tokyo for what is called a lecture. Most of those who have attended the sessions of Presidents' Clubs over the years have suggested when asked that the meetings are primarily social in nature.

The Presidents' Clubs list is at best a first cut at determining de facto keiretsu membership, as the participation of the automotive industry illustrates. Among vehicle makers, Toyota Motor Corp. belongs to the Mitsui Group, Mitsubishi Motors Corp. to the Mitsubishi Group and Nissan Motor Co., Ltd. to the Fuyo Group, yet Sakura Bank's influence over Toyota is limited, as is that of Fuji Bank over Nissan.

Beyond this trio, the situation becomes more complicated. Although Mazda Motor Corp. is not a member of the Sumitomo Group's Presidents' Club, it long has been considered a part of that keiretsu. Two decades ago, Mazda was on the receiving end of a well-publicized rescue orchestrated by Sumitomo Bank.4 Today, Sumitomo Bank remains the automaker's largest lender, followed by Sumitomo Trust & Banking Co., Ltd. Together, they are at least three times as important as the next-largest lender, Bank of Tokyo-Mitsubishi. Two of Mazda's 35 directors come from Sumitomo Bank.5

Yet Mazda's biggest shareholder is Ford Motor Co., whose 33.4 percent stake dwarfs the combined 8 percent or so holdings of Sumitomo Group firms. Moreover, a Ford representative heads the company, and four board members are from Detroit. Mazda and Ford also have links at the operating level within the increasingly global automotive industry.

Isuzu Motors Ltd. is a member of the Dai-Ichi Group's Presidents' Club, yet its relationship with General Motors Corp. resembles that of Mazda and Ford. The world's top vehicle builder owns 49 percent of Isuzu's stock, far ahead of runner-up Dai-Ichi Kangyo Bank, which has a mere 2.3 percent stake. DKB, however, is Isuzu's the largest lender, with roughly a 15 percent share.

Similarly, Daihatsu Motor Co., Ltd. is a member of the Sanwa Group, but otherwise its connection seems obscure, to say the least. The minicar maker is closely tied to Toyota — nominally a Mitsui Group member — which holds 51.1 percent of its shares. Sanwa Bank owns only 0.9 percent of the stock. Among other financial institutions, both Sumitomo Trust & Banking and Mitsui Mutual Life Insurance Co. are more prominent stockholders. None of these three financial institutions appears to be important as lenders. Daihatsu's 30-member board of directors does not include representatives from any other Sanwa Group company.

Honda Motor Co., Ltd. is regarded as independent of all automotive makers and banks, foreign and domestic. It belongs to no Presidents' Club. Yet its ties to the Mitsubishi Group are more than trivial. Its three largest stockholders are Mitsubishi Trust & Banking Corp., Bank of Tokyo-Mitsubishi and Tokio Marine & Fire Insurance Co., Ltd., in that order. Together, the three own 14.6 percent of Honda's stock. Bank of Tokyo-Mitsubishi and Mitsubishi Trust are Honda's largest lenders, and a BTM representative sits on the firm's 40-member board.

In short, the Presidents' Clubs list is not necessarily an accurate reflection of the links between automotive makers and their primary financial institutions. The Daihatsu example suggests that Presidents' Club membership may overstate the influence of the leading keiretsu, while the Honda example indicates that the clout of the Big Six extends to companies that are not part of a Presidents' Club. This uncertainty has led to a plethora of compilations of keiretsu.6 Yet these alternatives also make arbitrary choices as to who is in and who is out, with a lot of movement back and forth from year to year.

Presidents' Club criteria are, on balance, a very conservative measure of the impact of the Big Six on the Japanese economy, but even by that yardstick, they loom large (see Table 2). Excluding their bank and insurance company affiliates, in FY 1998, the Big Six accounted for 3.2 percent of total employment, down from 4 percent in FY 1988, and 11.5 percent of total sales versus 13.6 percent 10 years earlier.7 The relatively high proportion of sales is a result of the presence of one or more trading companies in each group, as is the declining share of employment and revenues over the past decade.
e--aaaaa- --

Table 2: Role of Big Six Keiretsu in Japan's Economy, March 31, 1999














Group Lending







Employees (thousands)







Share of Employment







Sales (trillions of yen)







Share of Economy







Source: Toyo Keizai Shinposha, Kigyo Keiretsu Soran (Tokyo: 2000), p. 23 and p. 40.

Among the Big Six, Mitsui, Mitsubishi and Sumitomo account for less of the economy than the other three major keiretsu. Nonetheless, their cohesion is greater, as evidenced by the fact that the percentage of each group's shares held by other members is higher. The same holds for the share of borrowing done from group companies.

Significantly, the Big Six do not necessarily have influence over Japan's economy proportionate to their share of economic activity. For decades, some respected experts have argued that the organizational structures of these groups as well as market forces severely restrict their ability to act as a unit.8 Not only are member firms forced by competitive factors to find the lowest source of supply or the best price for their products, even if the partners are outside the keiretsu, but some group companies also compete with each other.


Nonfinancial Keiretsu

Nonfinancial, or vertical, keiretsu are quite different. Their lead companies are not banks but nonfinancial firms, typically manufacturers, that have equity and other links to firms up and down the production chain. An automotive maker, for example, may own stock in both upstream suppliers and downstream distributors. Moreover, personnel and loans can flow from the lead company to these suppliers and distributors.

The equity and personnel ties alone may be far more significant than those that link horizontal keiretsu, but these companies also have important trading relationships. An automotive producer and a parts supplier may work together to develop new products, for example. Moreover, in contrast to horizontally linked companies, which may do relatively little business with each other, the chief customers of vertically linked companies might be each other.

Vertical keiretsu companies, in turn, may be part of a single financial grouping or, as in the cases of Toyota and Daihatsu, members of different ones. Some vertical keiretsu also take on a few of the trappings of horizontal keiretsu, diversifying into activities beyond their core businesses.

Nissan, for example, traditionally had close ties not only with Nissan Fire & Marine Insurance Co., Ltd. — a logical vertical keiretsu fit since one of the insurer's main activities is selling automobile insurance in connection with new vehicle purchases — but also with Nissan Mutual Life Insurance Co., whose ties with the car and truck industry were less straightforward. In this instance, the connection reflected Nissan Mutual's history as part of the prewar Nissan shinzaibatsu, an attempt by the vehicle builder to create a zaibatsu in the mold of Mitsui, Mitsubishi and Sumitomo. In any case, when Nissan Mutual failed in 1997 (see JEI Report No. 17B, May 2, 1997), the automaker, in less than robust health itself, refused to help its longtime partner.

Vertical keiretsu relationships are far too pervasive and complicated to be captured the same way that the Presidents' Clubs membership list does for horizontal keiretsu. Most major manufacturers and many retailers and railroad firms, for example, have relationships that can be described as a vertical keiretsu; each of these can involve hundreds of companies. Some sources count as many as 1,100 manufacturing groups.9 Such a total would include groups clustered around household names like Sony, Matsushita Electric Industrial Co., Ltd. (the maker of the Panasonic brand), Daiei, Inc. (Japan's biggest retailer), Seibu (a diversified retail and transportation concern) and many others less familiar. In some, but not all cases, vertical keiretsu companies are formally organized into groups of suppliers — for example, Mitsubishi Motors' Kashiwa-kai.

Hitachi, Ltd. currently is trying to collect royalties from 600 companies that use the Hitachi name or logo.10 That may be more companies than most vertical keiretsu encompass, but the number suggests how large such groups can become, especially if an inclusive definition is adopted.

Vertical keiretsu can evolve in three ways. First, and most obviously, existing members can establish a new group firm. Second, a group may realize that a major division or subsidiary of the lead company needs additional "space," perhaps so that it can solicit business from outside customers. This operation will be spun off, but it will maintain closer ties than, for example, an American company would under the circumstances. Eventually, the new company's stock might be publicly traded. Even if the business grows to global stature, it still may have a strong relationship with its former parent.

Denso Corp. (formerly Nippondenso Co., Ltd.) exemplifies this pattern. Founded as a division of Toyota in the 1930s, it was spun off as an independent company in 1949 when Toyota faced economic difficulties. In the 1950s, Nippondenso developed close ties with Robert Bosch GmbH, the big German parts maker.11 With the growth of Japan's automotive industry in the 1960s, 1970s and 1980s, Denso became a global company, supplying such products as speedometers to virtually every major car and truck maker. In 1998, according to Fortune magazine, Denso was the 305th-largest company in the world as measured by sales and the number-two parts supplier after Bosch.12 However, to this day, Denso's largest customer and shareholder — with a 24.6 percent share, virtually unchanged over several decades — is Toyota. Other Toyota Group companies, including Toyoda Automatic Loom Works, Ltd., out of which Toyota itself was formed, own another 7.1 percent. Bosch has a 5.2 percent interest.

The third way that vertical keiretsu evolve happens when a company, eager to obtain a secure source of supply or market for its products, "captures" other long-established companies. One such firm, Kato Shatai Co., Ltd., traces its origins to a wagon maker founded in 1901. It began the postwar era as a supplier of truck cabs and other products to Toyota, Nissan and Isuzu, among others. However, with the efforts of various automakers to bring parts production in-house in the booming early postwar years, Kato Shatai became part of the Mitsubishi Motors Group, which accounted for most of its sales. When Kato Shatai failed in 1976 in the aftermath of the first oil crisis, Mitsubishi Motors was obliged to bail it out and keep the plant operating. Kato Shatai emerged from bankruptcy in the 1980s.13

Table 3 and Table 4 show some of the firms forming vertical keiretsu around Toyota and Hitachi. The tables list only those companies identified as group firms in Kigyo Keiretsu Soran,14 one of the standard compilations of data on keiretsu. Companies whose stock is not publicly traded are excluded. The tables also omit some well-known names that apparently did not make the Kigyo Keiretsu Soran cut despite historical and other links to Toyota or Hitachi. For example, Hitachi Zosen Corp., a major shipbuilder, is missing. Even so, these groups are enormous. The Toyota Group's collective sales in FY 1998, including those of Toyota itself but not its brokerage and insurance affiliates, were ¥15.3 trillion ($140 billion at ¥110=$1.00). As such, Toyota Group sales were one-third those of the Dai-Ichi Group, the largest keiretsu as measured by revenues in Table 2.

Several aspects of vertical keiretsu stand out. First, the parent is likely to have what normally would be considered a controlling interest in group firms. Moreover, this heavyweight often is the main customer for members' products. If keiretsu "siblings" are considered, these tendencies become even more pronounced.

The relationships among vertical keiretsu firms were quite stable from the late 1980s to the late 1990s. Toyota has shown an almost uniform tendency to bring its affiliates more tightly into its orbit; Hitachi has exhibited a weaker pattern of movement in the other direction. Finally, while each group company typically specializes in an area related to a core competency, such as vehicle assembly, not all do. Toyota, for example, has a major interest in Kokusai Securities Co., Ltd., a big brokerage house.

The uniqueness of vertical keiretsu to Japan is easy to exaggerate. Manufacturers there have chosen an intermediate form of organization — nominally independent but actually linked companies — while American producers have opted either to keep parts and components operations in-house or to buy from truly independent suppliers. Vertical keiretsu seem quite similar to the in-house pattern widely observed in other industrial countries.

American companies, aware of the competitive advantages seemingly conferred by such vertical groupings, have moved to create relationships that are similar to keiretsu. Many U.S. producers have trimmed their supplier bases, apparently having learned from Japanese practice the value of working more closely with a smaller group of businesses.

Equally large numbers of firms — and, in some cases, the same companies — have moved in the opposite direction and are trying to give in-house operations their independence. General Motors exemplifies this trend. The world's biggest automotive maker spun off its Delphi Division in February 1999, creating with a single stroke the number-one parts manufacturer anywhere, a company far larger than Bosch and Denso. Delphi Automotive Systems Corp. now is aggressively going after non-GM business. Ironically, Delphi may have achieved nearly as much autonomy in one day as Denso has managed in more than half a century.

The distinctive aspect of Japan's vertical keiretsu — like their horizontal counterparts — is their durability. Just as GM did, Ford is moving to make its huge internal parts business, Visteon Automotive Systems, independent. Significantly, major components of Visteon began corporate life as separate concerns, only to be purchased several decades ago by Ford. This on-again, off-again pattern has no equivalent in the Japanese experience.


Causes Of Keiretsu

American and Japanese scholars alike long have asked why keiretsu developed in Japan and, with few exceptions, nowhere else. They also have wondered whether the effects of this corporate structure are favorable or unfavorable for the economy and Japan's social fabric.

Cultural Explanations - The most obvious explanation for keiretsu is that the groupings represent the extension to the corporate sphere of conventional Japanese preferences regarding the organization of personal and social life. Proponents of this view argue that, inasmuch as the Japanese are strongly attracted to groups, the organization of the corporate world along the same lines should not be surprising. With the most extreme interpretation derived from the mythical concept that the entire country acts as one corporation, the various keiretsu could be imagined to be subsidiaries of Japan Inc.

This argument has some glaring weaknesses, however. Zaibatsu were not ancient business organizations that persisted into the modern era but, rather, a business concept that evolved in the early phase of Japanese industrialization, which even now barely is 100 years old. The impetus for zaibatsu in the late 19th and early 20th centuries came from a state eager to modernize on the Western model; cultural preferences had little to do with their creation. Hence, it cannot be said that the predilection for keiretsu is deeply rooted in the Japanese psyche. Moreover, the conventional wisdom outside Japan regarding group behavior is based in part on American interpretations of wartime propaganda that urged all Japanese to "beat as a single heart." Wartime exhortations — especially the enemy's contemporaneous interpretations of the same — seem a weak basis on which to build a theory of social behavior and even less a means of understanding corporate behavior.

Market Imperfections - In this view, keiretsu arose because markets in early postwar Japan were riddled with imperfections. For instance, if firms with worthy investment projects and banks eager to lend money had a hard time finding each other and still more difficulty keeping track of one another once they struck a deal, financial keiretsu could act as intermediaries. Banks, as de facto members of the family, could be given greater access to borrowers' books and provide signals to other lenders, even those outside the keiretsu, that everything was satisfactory. In fact, banks could offer insurance against losses by non-keiretsu members by standing ready to bail out failed borrowers.

Agency Problems - Relationships between borrowers and lenders are fraught with inherent conflicts that are part of a class of difficulties known as "agency problems." In this case, keiretsu can solve problems that would exist even if markets functioned properly. For example, a lender naturally assumes that a borrower will accurately report on its business conditions only if things are going well.15 Therefore, the lender figures that the borrower usually is lying. Bringing the lender and the borrower inside the tent, so to speak, and giving each a stake in the other's profits might make the borrower more honest, thereby improving the quality of information and the likelihood that the economy's scarce capital is well used.

Similarly, a borrower normally would be almost indifferent to the choice between partial and complete default, yet a lender much prefers the former. Hence, from the standpoint of risk, the borrower might assign too little weight to the possibility of partial default relative to complete default, leading to choices that do not maximize the firm's present value. Again, cross-ownership of shares is a way of heading off decisions that, in the broader scheme of things, would be mistaken.

The implication of this argument is that the cross-ownership of shares at the heart of keiretsu improves economic performance. Regulations in the United States and elsewhere that limit or prohibit the ownership of stock by financial institutions are profoundly counterproductive in this view — as presumably would be efforts by keiretsu to dispose of cross-shareholdings.

Regulatory Determinism - This argument shifts the regulatory shoe to the other foot, asserting that Japanese regulations and tax policy have influenced corporate behavior, with adverse consequences for the economy.16 For example, the generous tax treatment given insurance companies has made life insurance an attractive means of saving. However, insurers traditionally have been very restricted in their choices of investments. They end up making the same kinds of investments and produce predictably similar returns. In this noncompetitive environment, insurers have little to offer corporate customers other than perhaps their keiretsu affiliation. To cement this relationship, insurers have tended to become stable shareholders in keiretsu companies.

Government policies that favor small firms through special tax treatment and specialized public-sector lending institutions are another example of how Tokyo has encouraged the growth of vertical keiretsu.17 Little wonder, then, that big businesses have a reason to set up legally separate small firms and to find ways to keep them under their control. Moreover, parents and subsidiaries can have different tax years, creating an incentive for income-switching from parent to subsidiary or vice versa. Today, newly formed operations must have the same tax year as the parent. That may explain why companies, once they have one or more subsidiaries, are reluctant to experiment with another corporate form.


Effects Of Keiretsu

Whatever the reasons for their emergence, keiretsu have had an impact on Japan's economic performance. Japanese analysts tend to believe that their country's extended run of spectacular growth originated partly in keiretsu. They argue that the long-term, patient approach to business relations implied by keiretsu of any kind makes Japanese firms more competitive. However, their argument is undermined to some extent by the reality that several of Japan's most competitive companies, Sony and Honda being prominent examples, have weak-to-nonexistent keiretsu ties.

At the same time, keiretsu have been linked to the protracted crisis in Japan's financial sector. In this view, banks, which are supposed to act as monitors of borrowers' financial health, have an incentive to be less than candid with other lenders if things are going so badly as to reflect on their own situation. Moreover, if a borrower requires a bailout, its lead bank may be in no position to provide one, thereby defaulting on the implicit contract said to be part of intra-keiretsu lending. In short, in the presence of keiretsu, troubled borrowers and weak banks lead to an environment in which a bank keeps its mouth shut and its fingers crossed, increasing the likelihood that a bad situation will get worse.

U.S. trade negotiators, armed with studies by American economists, sometimes have claimed that keiretsu are a potential trade barrier.18 Their argument in brief is that both vertical and horizontal keiretsu favor their members at the expense of outsiders, the latter of which, of course, includes American competitors.

While by no means do all economists on the two sides of the Pacific subscribe to the finding that keiretsu have a blanket exclusionary impact on outsiders — or any big-picture impact whatsoever — they are more likely to accept this conclusion when it is applied to specific industries. In this regard, insurance is a candidate for Exhibit A. As noted, insurers, whether writers of life or property and casualty policies, traditionally have had little reason to compete, implying that corporate customers have not felt compelled to shop around. For the huge amounts of insurance bought by businesses, a keiretsu insurer is as good as any and may be even better if it owns a block of stock in the company.19 Not surprisingly, insurance providers have tended to be members of the same keiretsu as the firms purchasing coverage.

Equally unsurprising has been the continued prominence of insurance in bilateral negotiations since the start of the 1990s (see JEI Report No. 12B, March 24, 2000). Washington has supported world-class U.S. competitors that have faced a multitude of problems in trying to crack the Japanese market. In recent years, keiretsu ties have assumed greater visibility on the White House's Japan trade agenda, in part because more overt barriers have been dismantled.


Recent Trends In Keiretsu

Another set of trends involving keiretsu has caught the attention of Japan watchers. For at least the last two years, the media in Japan as well as elsewhere have been reporting that cross-shareholding — a measure of the extent to which companies own shares in each other — is in rapid decline. If confirmed, these reports could mean, as some analysts already have claimed, that the whole keiretsu structure is crumbling.

Most of the accounts are based on reports of stock sales by financial institutions; normally, the net change in ownership is not cited. Moreover, cross-stockholding is an ambiguous concept. The reports do not describe with precision what is meant by the term or specify the threshold level, both of which are critical.20 For example, if two companies sell some of their stock in each other, the new level of cross-holdings may be lower than the level that triggers a report. Calculated cross-holdings between this pair would drop to zero, even though actual holdings may have declined only slightly.

The anecdotal evidence dovetails with Tokyo's position, which basically boils down to the argument that because keiretsu are disappearing, American negotiators should back off. In fact, based on the rationale for keiretsu and the demonstrated effects of this corporate structure, a rapid crumbling of the system would not be surprising. The regulations that encouraged the growth of keiretsu are on their way out, or are supposed to be. Moreover, Big Bang financial reforms are making it easier for firms to raise money without going to keiretsu banks, which increasingly are becoming the dinosaurs of Japan's financial services industry.

Ample indications also exist that nonfinancial firms' horizontal and vertical keiretsu are unraveling. For instance, every one of Japan's nine car and light truck builders except Toyota and Honda now has or soon will have partial foreign equity ownership and its accompanying financial links. In late March, Mitsubishi Motors, heretofore a stalwart member of the Mitsubishi Group, announced that it would link up with DaimlerChrysler AG. This move could threaten not only the vehicle maker's ties with Bank of Tokyo-Mitsubishi and Mitsubishi Steel Manufacturing Co., Ltd. but also with members of its own supplier network. Indeed, Nissan, as part of its restructuring under the watchful eye of Renault S.A., will dramatically cut its supplier numbers. Interestingly, Japan's main business daily, Nihon Keizai Shimbun, reported in March that Delphi is contemplating investing in one of the largest, Calsonic Corp.21

Banks themselves seem to be encouraging the collapse of keiretsu. In recent months, Japan's nine nationwide commercial banks have announced a series of mergers that, if fully consummated, will leave standing by mid-decade four large bank groups and a handful of second-tier institutions (see JEI Report No. 12B, March 24, 2000). Of the Big Six keiretsu banks, Sumitomo Bank and Sakura Bank (the lead bank for the Mitsui Group) will combine, as will Fuji Bank and Dai-Ichi Kangyo Bank. Sanwa Bank probably will gain companies for its group through a combination with the smaller Tokai Bank and Asahi Bank. Only Bank of Tokyo-Mitsubishi is not affected by merger fever. The blended keiretsu that emerge from this process may prove themselves totally incapable of acting in cohesive fashion. That prognosis could be especially applicable to longtime competitors Mitsui and Sumitomo, which have been fierce rivals since their zaibatsu days.

The Figure adopts a different measure of cross-shareholding. It also covers trends over a much longer period of time than most recent press accounts, which emphasize developments over a year or two. The data refer to the percentage of shares held by Presidents' Club members of the Big Six. This number gradually rose in the 1980s before peaking at 25.8 percent of all shares in FY 1993. Since then, it has dropped slightly, but, as of March 31, 1999, when the argument that keiretsu were collapsing already was in full fury, the figure still was higher than in FY 1983, the first year of this data series.

Other statistics actually show the role of the Big Six increasing in recent years. Between FY 1995 and FY 1998, the only two years for which information is available, the number of companies for which banks claimed a 20 percent or larger share of their borrowing rose from 1,690 to 1,956 out of a universe of 2,432 firms.22 These data would suggest that companies are becoming more dependent on a relatively small number of institutions, a trend that would be consistent with tighter keiretsu linkages. However, borrowing in general was extraordinarily weak during this period due to a "credit crunch." A decline in the ranks of banks also might have contributed to the finding.

Other evidence regarding the staying power of keiretsu is mixed as well. While some banks have disposed of shares in keiretsu companies and vice versa and the planned bank mergers point to weakened keiretsu cohesiveness, recent moves by the Mitsui, Mitsubishi and Fuji Groups to more closely align their commercial bank, trust bank and nonlife and life insurance units as a means of becoming financial services supermarkets suggest that bankers believe that there still is life in their organizations. For example, as banks prepare for the sale of insurance at their branches beginning April 1, 2001, a result of the Big Bang, they are tapping their traditional keiretsu partners for products.

Moreover, the pending bank consolidations are leading to secondary mergers among keiretsu firms. Keiretsu insurers are following their lead banks in announcing combinations (see JEI Report No. 11B, March 17, 2000). For instance, Mitsui Marine & Fire Insurance Co., Ltd. abandoned a planned merger with Nippon Fire & Marine Insurance Co., Ltd., a member of the Dai-Ichi Group, and Koa Fire & Marine Insurance Co., Ltd., an independent, after Sakura Bank announced its linkup with Sumitomo Bank. Instead, Mitsui Marine & Fire will merge with Sumitomo Marine & Fire Co., Ltd. effective October 1, 2001. Similarly, reflecting the proposed merger of Fuji Bank and Dai-Ichi Kangyo Bank, two of their respective insurers, Yasuda Fire & Marine Insurance Co., Ltd. and Fukoku Mutual Life Insurance Co., will team up. The Big Bang might have been expected to lead to the exact opposite result, in which keiretsu companies looked for the best partners regardless of affiliation.

The question remains: Will keiretsu crumble into dust, additional victims of progress and financial liberalization? Perhaps — but their demise often has been predicted before. A 1990 article was entitled (in translation) "Separation from banks accelerates: What will happen to kigyo keiretsu?"23 The implication was that Toyo Keizai Shinposha, the venerable publisher of Kigyo Keiretsu Soran, might have to discontinue its long-running summary of keiretsu activity. Notably, its 2000 edition is as thick as ever.



The developing picture of financial keiretsu is one of considerable consolidation across and within groups, suggesting the likely emergence sometime in the middle of this decade of four perhaps truly mammoth conglomerates offering a wide array of financial services. However, these competitors probably will be less closely linked to nonfinancial members of historical keiretsu. There will be too many traditional rivals within any one group to make cooperation work. In other words, economic power will be more concentrated or less so, depending on one's point of view.

Nonfinancial firms will be subject to other forces as well. Those that ally with foreign companies, as most automakers have, may be subject to intense pressure to loosen ties to their suppliers. These businesses' vertical keiretsu links will weaken, and some suppliers even may fall into an orbit around a foreign keiretsu wannabe like Delphi. But those companies not forging alliances with foreign corporations may move to strengthen ties within their group to prevent suppliers from going the same way — a tactic that Toyota appears to have adopted.

Finally, the emergence of brand-new links modeled on keiretsu — or even zaibatsu — cannot be ruled out. Some observers see natural zaibatsu structures in the emergence of Internet firms, a perception that grew stronger after Internet giant Softbank Corp., as part of a consortium, made the successful bid for Nippon Credit Bank, Ltd., one of Japan's traditional long-term credit banks and the core of a keiretsu of its own. Softbank already has interests in many Internet-based financial services companies, most of them American.

As with so many other issues surrounding the Internet, the viability of such Net-batsu remains very much up in the air. Such a possibility suggests that the American fascination with Japanese corporate groups will remain strong more than a half-century after occupation officials tried to root them out.

Kanako Yamada provided research assistance.

The views expressed in this report are those of the author
and do not necessarily represent those of the Japan Economic Institute

Top aaaa Issue Index aaaa 2000 Archive Index aaaa Subscriber Area aaaa Home


1aa This period is described in detail by one of the American participants in Eleanor Hadley, Antitrust in Japan (Princeton, New Jersey: Princeton University Press, 1970). Return to Text

2aa This surge in entrepreneurship is discussed in its historical context and compared with contemporary Japan in Douglas Ostrom, "The Search For New Corporate Superstars: Japanese Firm Mobility In The 1990s," JEI Report No. 12A, March 26, 1999, particularly pp. 9-11. Return to Text

3aa Hideaki Miyajima, "The Transformation of Zaibatsu to Postwar Corporate Groups — From Hierarchically Integrated Groups to Horizontally Integrated Groups," Journal of the Japanese and International Economies, No. 8, September 1994, pp. 300-302. Return to Text

4aa See Richard Pascale and Thomas Rohlen, "The Mazda Turnaround," Journal of Japanese Studies, Summer 1983, pp. 219-263. Return to Text

5aa The data in this section are from Toyo Keizai Shinposha, Kigyo Keiretsu Soran (Tokyo: various years). Unless otherwise stated, the data refer to the fiscal year ending March 31, 1999, the most recent period for which information is available. Return to Text

6aa The best-known English-language listing is Industrial Groupings in Japan — The Anatomy of Keiretsu, published biennially by Dodwell Marketing Consultants, Tokyo. Return to Text

7aa The totals are smaller than the sums across the six groups because of overlapping group membership. Return to Text

8aa See Douglas Ostrom, "Keiretsu And Other Large Corporate Groups In Japan," JEI Report No. 2A, January 12, 1990, p. 12, and the sources cited therein, and Douglas Ostrom, "Competitive Conditions And Competition Policy In The Japanese Economy," JEI Report No. 9A, March 10, 1995, pp. 4-7. Return to Text

9aa One such source is Toyo Keizai Shinposha, Kigyo Gurupu (Tokyo: various years), which is discussed in David E. Weinstein, "Foreign Direct Investment and Keiretsu: Rethinking U.S. and Japanese Policy." Paper presented to the Japan Economic Seminar, February 17, 1996, p. 22 and Table 4. Return to Text

10aa "'Hitachi' Shamei Shiho-ryo: Gurupu 600-sha-kara Choshu (Charges for Using 'Hitachi' Name: Collections From 600 Companies)," Nihon Keizai Shimbun (American edition), March 20, 2000, p. 1. Return to Text

11aa For details of Denso's history, see Konosuke Odaka, Keinosuke Ono and Fumihiko Adachi, The Automobile Industry in Japan: A Study of Ancillary Firm Development (Oxford, England: Oxford University Press, 1988), pp. 154-169. Return to Text

12aa "1999 Fortune Global 500." Available at www. Return to Text

13aa This description of Kato Shatai is summarized from Michael J. Smitka, Competitive Ties: Subcontracting in the Japanese Automobile Industry (New York, New York: Columbia University Press, 1991), pp. 115-134. Return to Text

14aa Published annually by Toyo Keizai Shinposha, Tokyo. Return to Text

15aa David Flath, "Keiretsu shareholding ties: Antitrust issues," Contemporary Economic Policy, XXII, January 1994, pp. 24-36. Return to Text

16aa Weinstein, op. cit., especially pp. 18-27. Return to Text

17aa Weinstein, op. cit. Return to Text

18aa See Douglas Ostrom, "The Structure Of Japan's Imports: Causes And Consequences," JEI Report No. 22A, June 14, 1991, pp. 7-9, and the sources cited therein. Return to Text

19aa See Douglas Ostrom, "Japan's Sleeping Insurance Giants: Roused And Ready?" JEI Report No. 17A, May 1, 1992, pp. 12-14. Return to Text

20aa One of the most comprehensive reports is Hideaki Inoue, Recent Developments in Cross-Shareholding (NLI Research Institute Working Paper No. 121) (Tokyo: Nippon Life Insurance Co., 1998), p. 17. Return to Text

21aa "Karusonikku-ni Shusshi Kento (Calsonic investment contemplated)," Nihon Keizai Shimbun (American edition), March 6, 2000, p. 11. Return to Text

22aa Toyo Keizai Shinposha, Kigyo Keiretsu Soran (Tokyo: 2000), p. 66. Return to Text

23aa Toyo Keizai Shinposha, Kigyo Keiretsu Soran (Tokyo: 1990), p. 26. Return to Text

The views expressed in this report are those of the author
and do not necessarily represent those of the Japan Economic Institute

Top aaaa Issue Index aaaa 2000 Archive Index aaaa Subscriber Area aaaa Home