2003/04 Research & Review

The Implications of Modular Architecture

Faculty Fellow, RIETI

If the 1980s is an era characterized by the downfall of an American economic model and the rise of a Japanese economic model, then from 1990 onwards is an era in which the Japanese model came apart and the American model revived. This tendency, though subject to some nuance, is continuing to date. As to the fundamental cause of the failure of the Japanese economic system and clues for restoring the nation's industry, however, opinions differ substantially. I believe that one reason behind Japan's stagnation lies in the changes the modern industries have undergone. My research is based on this understanding.

Modularization of Industrial Architecture

The revival of the US economy in the 1990s has been primarily supported by development of the information and the communication industries. With venture capitalists providing loans and funds, a number of entrepreneurial firms have emerged especially in areas in and around California's Silicon Valley. The consequence of this is the drastic transformation of industrial structure. Specifically, the computer industry, which used to be a monopoly market dominated by International Business Machine Corp. (IBM), was divided into various sub-industries, in each of which small and midsize entrepreneurial firms embarked on diverse research and development (R&D) activities. Through this process, the aggregate market value of the computer-related industry has sharply increased. Among a range of factors behind such a drastic structural change, the modularization of architectures, such as in computers and the Internet, has been the most important.

Simply put, the modularization of architectures is the minimizing of inter-component dependency when designing a complex product system so that each component (or modular product) can be independently developed, and then combined afterwards to form the product system. Personal computers (PCs) are a ready example of a product system developed in such a manner. Also, communication technologies like the Internet can be cited as another example of modularization in a sense that they are formed by component technologies of various layers. Modularity brings the following benefits: (1) new products can be developed in quick succession while reusing the same architecture, and (2) respective component developers can work independently with little worry over coordination with those developing other components, thereby realizing the benefit of specialization. Many entrepreneurial firms in Silicon Valley have grown by developing and producing modular parts of a product system, not by developing and producing a stand-alone product system. The presence of such a modular architecture has contributed to the drastic change of landscape in the US computer industry.

Theoretical Elucidation of Modularization

Benefits brought forth by modularization are not confined to those described above. Looking from an integral viewpoint, Baldwin and Clark analyzed how modularity generates value. They pointed out that the fruits of R&D activities can only be regarded as real options when their merits exceed those of conventional products. For instance, by modularizing a system design and aggregating the achievements of numerous small R&D modules, modularity generates far greater results (the effect of splitting operator) than those born out of R&D activities on non-modular products. Also, as many companies or inter-company divisions compete in the development of a single module, the value of a modularized product system increases rapidly (the effect of substituting operator).

Of course, the modularization of a product system that has complex inter-component dependency incurs costs, so it is not a case of "the more modularized the better". Schaefer and Takizawa have shown that the optimal extent to which a product system should be modularized depends on the cost of internal coordination within each module development team (the cost of within-team coordination) and the cost of coordination with other development teams (the cost of across-team coordination). More specifically, the higher the cost of within-team coordination or the lower the cost of across-team coordination, the finer the optimum extent of modularization becomes. In a situation where the cost of across-team coordination can be lowered by investing to create an efficient communication network (this can be regarded as information technology investment in today's context), the higher the cost of within-team coordination, the more a company should spend on IT investment to lower the cost of across-team coordination. The results of this analysis correspond to the result of an empirical analysis, which showed that the size of an American firm has become smaller in accordance with its increase in IT investment. The findings are also consistent with the fact that Silicon Valley companies, which feature high mobility of workers and a substantial degree of information sharing amongst different firms, tend to be smaller in scale than their counterparts in Route 128 in Massachusetts, an area which has the high concentration of IT-related industries but features lower mobility of workers and little information sharing among firms.

Furthermore, Aoki and Takizawa attempted to find what kind of a cross-team information sharing is desirable among firms developing different parts for a single product system. Based on the findings, we concluded that an information system that combines decentralized information sharing and information encapsulation is the optimal choice in developing a modularized product system. This can be interpreted as reflecting what Saxenian described as an information system unique to Silicon Valley.

Modularization and Venture Capitalists

Another significant aspect of the ongoing modularization is that entrepreneurs engaged in modularized R&D activities are receiving financial supports from angels and venture capitalists. Apart from acting as an intermediary for information exchange among entrepreneurs, venture capitalists are playing an important role in the governance of entrepreneurial firms. In the outset, venture capitalists would rarely provide the full amount of capital required to carry out a project. Instead, they would offer seed money, a fractional fund just enough to start up, and then provide additional funds in accordance with the performance and progress of the project.

When multiple entrepreneurial firms compete in a tournament under such circumstances, the aforementioned "effect of substituting operator" would emerge. As a result of multiple entrepreneurs devoting their efforts to the tournament, there is an inevitable duplication of costs for the society as a whole. However, by explicitly analyzing incentives for entrepreneurs, Aoki has shown that the mechanism of the tournament, despite the duplication of costs, could effectively function if the prize for the winner in the tournament is sufficiently large and the venture capitalists' ability to determine the winner based on the value of R&D results is sufficiently high.

By endogenizing the number of participants in the tournament as a variable, Aoki and Takizawa examined how these participants are affected by marketing and technological uncertainties. We have found that this tournament mechanism can entice greater efforts from the participants, thus functioning best when technological uncertainty is high and marketing uncertainty is low. It is well known that a number of electronic commerce (e-commerce) firms sprouted during the dot.com bubble period but that many of them failed, causing the bursting of the bubble. The erroneous expectation concerning the profitability of dot.com companies may be primarily responsible for the crash of the bubble. At the same time, however, certain factors characteristic of e-commerce - a relatively low start-up cost and a great number of participants as its consequence, and the combination of high marketing uncertainty and low technological uncertainty - are believed to have played a role.

As mentioned above, the link between entrepreneur and venture capitalist is a relational one, starting with the venture capitalist's offering of seed money and followed by additional financing in accordance with the progress of the project. Their relationship is primarily defined by a venture capital contract, the contents of which are quite interesting. Among series of contracts formed in the several stages of business development, those in an early stage typically include the vesting of stock options to an entrepreneur as a means to elicit stronger commitment from them, explicitly noting that the business idea of the entrepreneur is a precious asset. Meanwhile, contracts formed in a late stage tend to provide greater controlling authority to the venture capitalist because professional management becomes important; for instance, to decide on the timing of initial public offering or divestiture. It is also observed that a venture capitalist's controlling authority tends to increase when a project is not doing well while the controlling authority of an entrepreneur becomes stronger when the project proceeds successfully. The conventional approach to understand property rights does not fully explain these characteristics demonstrated by venture capital contracts, which thus present an interesting subject for analysis in the field of corporate theory.

Recent Theoretical Development Concerning Industrial Dynamics

The analysis of modularization mechanisms introduced here does not detail how the modularization of a product system and subsequent transformation of industrial structure have occurred. More recent researches attempt to explain this from an increasingly dynamic perspective by incorporating the analysis results of modularization mechanisms. For instance, Christensen and others presents a hypothesis as follows:

In a situation in which customers are dissatisfied with the function of products available on the market, integration becomes a dominant trend in an industry as companies try to improve the quality of products while maintaining complex interdependency. However, because the speed at which companies improve the quality of products is greater than the speed at which customers absorb and digest the new technologies, the relation between customers and products eventually reverse with customers more than satisfied with the functions of products on the market. At this stage, they contend, competition enters a new phase with focus shifting on to speed and prices, and the modularization of products begins.

As Christensen explains in separate works, a shift from integration to non-integration is related to phenomena in which leading companies often decline when they face destructive innovation. If the hypotheses by Christensen and others are correct, the reason why Japanese companies lagged behind their foreign rivals in the field of information technology (IT) could be attributed both to the absence of powerful startup firms, and to integrated companies' failure to adapt to the quick development of industrial dynamics. Exposed here again is the weakness of Japanese companies, that is, their inability to implement drastic organizational change and, in particular, break internal constraints developed over the years.

Intellectual Property Right and Industrial Dynamics

Christensen and others have not provided micro foundation to their hypotheses. In proceeding on more detailed analysis of the process of industrial dynamics, its relationship with intellectual property rights must be examined. Acknowledging that intellectual property rights are generally of little efficacy, Teece examined what kinds of strategies enable innovators to secure profits from their innovation. The effective power of intellectual property rights is generally weak in the IT industry. Under such situations, the optimal strategy for an innovator is to tie up with a company that has assets complementary to the innovation. This explains why many Silicon Valley entrepreneurs prefer to be acquired by a leading company instead of trying to make an IPO on their own, and why companies such as Cisco Systems, Inc. try to absorb entrepreneurial firms in acquisition and development (A&D) strategies.

As a premise for these phenomena, however, the modular architecture of a larger product system needs to be disclosed to the public and there must be a system to enable anyone to make changes to the system. In other words, the modular architecture needs to become the "commons" as defined by Lessig. As Lessig persuasively explains, attempts are being made today to recapture control over technologies that have already become the commons, using intellectual property rights as leverage. It should be noted, however, that such attempts may end up narrowing down the possibility of innovation.

Research efforts toward explaining industrial dynamics by incorporating the analysis results of modularization mechanisms have only just begun. How this process is affected by the extent of intellectual property right protection is a very interesting question, but this too is a subject of future study.

Modularization and Other Implications

What has been described above tried to explain the mechanism of modularization bearing in mind the drastic structural changes in the United States. But the findings on the benefits of modularization discussed here also have implications to other areas. In the 1990s, many countries experienced an economic crisis and began to reform fiscal structure. Policies commonly observed in many of these reforms are the reinforcement of centralized control over budget allocations and the introduction of an independent agency system together with a system for ex-post evaluation. Such a trend can be perceived to be a kind of modularization within the public sector. Each independent agency is related to the rest of the government through its fulfillment of explicit goals but decision-making on the allocation of resources within the agency is encapsulated.

The biggest difference between public services and those offered by the private sector is that a monopoly is justified in the former whereas companies in the leading position in the private sector are replaced from time to time through continuous innovations. This difference is believed to be responsible for impairing the public sector's responsiveness to changes in environment. By modularizing government activities wherever possible, the public sector, at least partially, would be exposed to the pressure of innovation. The introduction of the independent agency and ex-post evaluation system is to adopt the wisdom of the private sector in the government sector.

For bibliography, please refer to RIETI Discussion Paper 03-E-009 ("Property Rights and the New Institutional Arrangement for Product Innovation in Silicon Valley").

>> Original text in Japanese

July 14, 2003

Article(s) by this author