The Promise of Virtual Keiretsus

The following discussion took place at a seminar sponsored by the Aspen Institute (www.aspeninst.org) for executives of leading software companies. We are grateful to the Aspen Institute for granting us permission to post this selection.

The most successful intranets tend to depend upon a community of openness, mutual commitment, and trust, which help an organization respond more flexibly and rapidly to external circumstances. It turns out that an organization based on such "soft values" can function more effectively and efficiently than most rule-driven, hierarchical systems of management, which tend to have higher transaction costs, less plentiful and diverse market intelligence, and slower response times. Now that many companies have demonstrated the benefits of networking within the firm, there is increasing interest in exploiting similar benefits that might come from networking among firms.

The Japanese Keiretsu

One of the most cited models for such interfirm collaboration is the Japanese keiretsu, a network of businesses that voluntarily enter into long-term relationships to benefit all of its members. The idea is to use trust and long-term commitment among partners to reduce the higher transactions costs that prevail when players have only episodic market relationships with each other. As Francis Fukuyama makes clear in his 1995 book, Trust: The Social Virtues and the Creation of Prosperity, a system of reciprocal moral obligation among firms can be a highly efficient way of achieving long-term competitive advantage:

"It is the long-term stability of the obligational relationship that is important: both contractors [in a deal] can invest and plan for the future knowing that the other will not jump ship if a third party offering a somewhat better price were to come along. They will, moreover, waste less time haggling over prices for any deal: if one party feels it got a less than optimal price or even suffered a loss in the short run, it knows that its partner will be willing to make this up at a later point."

Furthermore, Fukuyama points out, a firm affiliated with a keiretsu is likely to obtain superior market intelligence about its customers and competitors. It can undertake riskier ventures or long-term investments that may not yield returns until far into the future. It may be able to borrow money at lower real rates of return and enjoy more stabilized revenue streams. And it may reap reputational advantages from affiliating with other keiretsu members. By having relationships in which "'the major type of mutual acts is consensus/inducement-oriented,' and in which the actors have some kind of continuing yet informal relationship with one another," keiretsus can help a firm "achieve the savings in transaction costs of large organizations, while retaining the savings in overhead and administrative costs of large organizations."

The intriguing issue raised by Japanese keiretsus is whether similar sorts of long-term, trust-based relationships can be established through electronic networking. Is it possible to create "virtual keiretsus" that use relationships of reciprocal moral obligation to achieve competitive advantage?

The question is particularly significant because "no one has the capability of delivering a full range of services over the Internet without someone else," as Morton Meyerson points out. "Everyone has to partner." Furthermore, the speed with which technologies and markets change requires the major players to be able to play on many fronts simultaneously, and to mobilize rapidly to seize new opportunities. This requires firms to have ready access to diverse expertise and resources. Hence the great interest in finding new ways to coordinate diverse disciplines and companies more effectively, particularly internationally.

Interfirm relationships based on trust are also attractive because they can yield superior products--precisely because more room is sanctioned for learning and exploration. "A firm may not want to contract for a product with tight specifications because it will only get the product it asks for--and not something better, which might result if it were working within a relationship of trust with its vendor," noted Hal Varian of UC Berkeley. "The vendor would then know the contractor's goals and be willing to share the risk of failure, knowing that there is an upside in producing a better system or product."

How to Nurture Trust Among Firms in a Virtual Keiretsu

In imagining a virtual keiretsu, Morton Meyerson envisions it as "a neural network that is round rather than vertical, and has independent learning pieces that are interconnected and learn from each other. But the tricky thing is getting them to work together. You have to work very carefully to build relationships and trust," he said. "But if you leave it at this level, you won't get to a higher level of functioning." Rejecting the idea that legal contracts can facilitate this process, Meyerson believes that the interests of all participating parties must be closely aligned for the virtual keiretsu to work. This alignment of interests may require partnerships with common ownership, he said.

What makes the virtual keiretsu so problematic, concedes Meyerson, is "the abject fear of loss of control by the people involved. These feelings are so powerful and frightening that people will transmit the fear to their partners in the relationship." Overcoming this problem is difficult "because these feelings go to the core of a person," said Meyerson. "People won't let go to other people." The idea of entering into a new type of community through networking also "runs against our American ethic of individualism," he added.

One of the foremost challenges may be to devise new organizational and legal structures to foster trust. "There is a lack of serious research about how trust within and between corporations gets built," noted John Seely Brown of Xerox PARC. "How do you structure a business process that has enough elbow room to actually help the growth of trust within the firm? Typically, we do almost the opposite," Brown pointed out: "We try to nail down business processes so that they are so crisp, so precise, that even if you execute them satisfactorily, there is no chance to build trust among the people that execute them."

For virtual keiretsus to work, then, we may need to find new ways "to structure contracts, not for monitoring and control, but for the growth of trust and knowledge," said Brown. "Some economists and political scientists are currently exploring this challenge--how to give people a lot of room to fail, but also a lot of room to excel. Transposing that intra-organizational dynamic to relationships between firms needs greater attention." If such trusting collaborations between firms were difficult to achieve in the past (due largely to the high transaction costs cited by Ronald Coase above), the new technologies are significantly altering that equation by creating cheap, versatile ways of interacting with outside vendors and partners.

The best way to build trust, avers Eric Schmidt of Sun Microsystems, is to have clear and explicit boundaries for different functions that are agreed to by everyone involved. Keiretsu-like collaboration via electronic networking already occurs in such sophisticated processes as aircraft production. Boeing, when it produces a new aircraft, "is in effect acting as a systems integrator whose main business is to organize the activities of a host of independent subcontractors that do much of the actual manufacturing of the airplane," writes Fukuyama. In business endeavors where boundaries are not easily set, however, such as in the software development business, it can be much harder to integrate diverse players, nurture trust and form an ongoing virtual keiretsu.

Forging a successful virtual keiretsu may be especially complex, noted Morton Meyerson, because "you're not talking about the alignment of interests between two different enterprises, but the alignment of interests of internal units within an enterprise, which must then be coordinated with another firm's internal units, which themselves must be aligned within their firm--while trying to align the overall interests of the two enterprises. You've got so many complex points of alignment, it's unbelievable."

Negotiating the tension that exists between boundary-setting and openness, and between accountability and trust, is difficult. It is well-expressed by two contrasting lines of poetry from Robert Frost's "Mending Wall": "Good fences make good neighbors," and "Something there is that doesn't love a wall." Without sufficient internal structure and boundaries, an organization will not be effective and productive. But those boundaries must not be so rigid or pervasive as to stifle trust and creativity. The difficulty in forming virtual keiretsus, said John Seely Brown, is finding the right level of structure to manage "the rich interplay between ensembles of communities of practice, which form the social fabric."

William H. Janeway of E.M. Warburg, Pincus & Co. described a successful, noncoercive keiretsu comprised of about 90 companies in which his firm owns a significant stake (between 15 percent and 85 percent, for an aggregate value of $6.5 billion). Warburg, Pincus has sought to create a common economic interest among its operating partners based on trust and need, not mandates. "The idea is that, between each of our partners, there should be preferred relationships. This has evolved slowly and haphazardly, and that is absolutely, necessarily the case. It should be that way."

Through loose affiliations with each other, Warburg, Pincus partners "can leverage themselves into better positions through access to information, resources, and other parallel partnerships within the network of Warburg, Pincus enterprises," explained Janeway. "But if use of those resources is required, then immediately the relationship of trust is compromised and the benefits do not materialize." It is important to note that a multipoint network of self-evolving, peer-to-peer relationships does not happen automatically, said Janeway. It is a matter of creating new linkages and then educating its members on "how they can benefit from using the informational infrastructure for sharing all those things that you ought to know."

Why the Keiretsu Model May Be Misleading

Although the Japanese keiretsu offers an intriguing analogy for interfirm collaboration via networks, it may have some inherent limitations. The keiretsu comes out of a culture that is quite distinctive, after all, and not readily transferrable to organizations in Western culture. For example, Japanese keiretsus are intensely hierarchical, and are led by a strong company that essentially intervenes in the internal affairs of other companies, to the extent of training its members how to perform certain functions. But virtual keiretsus, as discussed here, involve mutual collaboration, the absence of full control, and the leveraging of trust.