|
By Masaaki Kotabe, Xavier Martin, and Hiroshi Domoto Illustrations by Gordon Studer
Of course, these relationships must be built to last. For what is effective in long-established relationships may not prove effective in newly-established ones. It takes time to develop the familiarity and expertise required for each partner to know when and how to draw on the other’s resources, or to contribute resources. As two firms sustain a business relationship over time, they develop a highly idiosyncratic joint understanding that allows for uniquely efficient communication. These relation-specific skills or relation-specific assets make ongoing collaboration more effective. Many firms held up as examples of successful knowledge management have developed comparatively long-lasting supplier links. But research has shed little light on whether and how firms with shorter links may benefit from knowledge-intensive sourcing, and on how the benefits of knowledge transfer vary with link duration. And while past studies show that buyers benefit when suppliers are intensively and durably involved in knowledge exchange, it is less clear under what conditions this improves suppliers’ operational performance. So we set out to examine the connection between knowledge transfer and link duration, to learn what it takes to enhance supplier performance. For our purposes, supplier performance was defined as a combination of product development efficiency, process improvements, quality conformity, and short lead-time. To do so, we developed several hypotheses, surveyed U.S. and Japanese automotive suppliers, and ran the data through a multivariate model.
Exchanges of Knowledge We analyzed two forms of knowledge transfer: technical exchanges and technology transfer. Conceptually, these two forms of exchange differ in the scope and level of the knowledge involved. A technique consists of discrete know-how required to solve a particular operational problem. So technical communications pertain to the relatively narrow and simple informational resources necessary to handle engineering issues. By contrast, a technology is a broader body of knowledge encompassing a set of related techniques, methods, and designs applicable to an entire class of problems. Its sharing or transfer involves higher-level capabilities.
The associations between technology and capability, and between technique and resource, add meaning to the distinction between technical and technological knowledge. Technical knowledge, consisting of narrower and more independent pieces of information, is a form of resources. Compared to discrete resources, capabilities are higher-order, more complex sets of routines with broader applications, and are harder to develop and mold. The coordination required for exchanging small-scale technical knowledge is typically simple. Arranging regular meetings or long-term personnel visits, for example, is straightforward if it involves autonomous individuals or small work units. As technical information tends to be explicit or at least codifiable, its exchange is a matter of verbal or written communication. By contrast, technology transfer involves a greater scope of activities and higher-level organizing principles. It requires extensive and dedicated coordination, as large and functionally diverse groups interact both within and across firms for sustained periods of time. This renders technology transfer particularly costly.
Four Hypotheses Past studies have argued that small-scale exchanges of technical information help improve the buyer’s performance. Suppliers likewise stand to benefit when the partners steadily share technical knowledge to solve problems and enhance products and processes. Therefore, as Hypothesis 1 suggests: The more technical exchanges between the buyer and the supplier, the higher the supplier performance improvement relative to two to three years earlier. Knowledge transfer, by contrast, requires larger-scale commitments of time and groups of experts. Still, projects that allow one partner to access or replicate complete technological capabilities of the other partner, when properly implemented, enable a more efficient division of labor, and distinct improvements in technological competence throughout the industry chain. That leads to Hypothesis 2: The more technology transfer between the buyer and the supplier, the higher the supplier performance improvement relative to two to three years earlier. The benefits of a long prior relationship stand to be larger yet when it comes to higher-level technology transfer. Technology transfer requires diverse functions of the supplier and the buyer to interact over multiple issues simultaneously. Under these circumstances, the benefits of having had the time to develop more relation-specific assets become all the more salient. That leads us to Hypothesis 4: The positive association between technology transfer and supplier performance improvement becomes stronger as link duration increases.
Testing Methodology To test our hypotheses, we examined
buyer-supplier relationships in the U.S. and Japanese automotive
industries. We
used well-established
industry directories to draw random samples. We collected the data
in two stages, first writing to prospective respondent firms to ascertain
their willingness to cooperate, and then sending questionnaires to
the individuals named by the firms. We asked a series of questions
about the respondent firms’ relationships with their main customer
(an automobile assembler, or for second-tier respondents, another
automotive component manufacturer). This yielded satisfactory response
rates for this type of research. Ultimately, our sample for analysis
consisted of 97 questionnaires, for a 24.3 percent response rate.
In Japan, we received 105 usable responses to the 577 supplier companies
that we solicited, or 18.2 percent. In defining both samples, we
excluded subsidiaries where an automotive assembler was a major shareholder,
as well as suppliers owned by foreign firms in each market. The responses
cover a wide range of products and firm sizes. On average, the U.S.
and Japanese suppliers were similar in annual sales (about $440 million)
and in total employment (3,100 to 4,500, not statistically different).
And most of the executives that responded held upper-management positions.
In the U.S. and Japan respectively, they had 4.66 and 5.58 years
of experience in their current positions and 14.01 and 19.90 years
with the same company. This exceeds the two to three year timeframe
for measuring our dependent variable, thus validating our research
design. We ran the tests for U.S. and Japanese samples, respectively. And then we analyzed the results to isolate the effects of variables such as Link Duration. The first analysis contains a base model without interaction terms. The second includes a single interaction term, between Link Duration and Technical Exchanges. The third contains a single interaction term, between Link Duration and Technology Transfer. The fourth analysis includes both interaction terms.
U.S. Results For the U.S. sample, the results support Hypotheses 1 and 4, but do not support Hypotheses 2 and 3. Simple technical exchanges can enhance supplier performance, and this effect is independent of whether a buyer and supplier have established familiarity through a long-established relationship. But we found that long-established links do not promote performance improvement by themselves. (However, they can help when combined with active technology transfer practices.) Meanwhile, technology transfer has no independent main effect but its combination with longer link duration is beneficial. Further analysis provided insight into the conditions under which technology transfer is likely to be beneficial. For example, the overall effect of Technology Transfer could be negative if Link Duration were very low, and becomes positive if Link Duration exceeds approximately five years. Interestingly, five years is approximately the length of a product design cycle in the U.S. automotive industry. Extrapolating this result suggests a distinct challenge for just-formed relationships: Premature use of technology transfer may harm supplier performance. Conversely, the payoff of transferring technology is particularly high in links of very long duration.
Common Ground The most similar results pertain to the significance and direction of the interaction effects. First, link duration does not moderate the effects of technical exchanges. Second, and more important yet, the effect of technology transfer increases with link duration, as per Hypothesis 4. The pattern of these results is broadly consistent with former Kyoto University researcher Banri Asanuma’s arguments whereby buyers should initiate supplier relationships with relatively simple tasks, and subsequently undertake more ambitious joint or delegated projects as the relationship matures. This argument has been generalized outside Japan based on Japanese assemblers’ behavior. Our results suggest that some substantive differences should nevertheless be taken into account when generalizing and implementing this recommendation.
In the U.S. sample, there is evidence that smaller-scale technical exchanges promote supplier performance improvement. Our tests also suggest that the distinction between medium- and long-duration links is more relevant in the U.S. than in Japan, where the primary distinction is between short links and longer links. However, no significant effect of technical exchanges is found in Japan. The pattern of results for Japan suggests that immediate payoff from technical exchanges may be elusive (relative to Japanese competition with longer-established links), while technology transfer is most beneficial once the sourcing relationship has been in place for a moderate period. Thus, failing to reconduct a relationship past the first purchasing cycle – or after subsequent cycles – entails a substantial opportunity cost. This may explain why Japanese buyers have long been described as comparatively reluctant to change suppliers (and vice-versa). Why do technical exchanges pay off in the U.S. sample but not in the Japanese sample? It is theoretically possible that U.S. firms are inherently more efficient at sharing such explicit knowledge. But prior research on the automotive industry does not support this view. A more plausible explanation is that U.S. buyers and suppliers have recently increased their commitment to joint problem solving. It is therefore comparatively easy to find technical exchange opportunities that enhance performance in the U.S., whereas Japanese firms, having long exploited this practice, have less to gain at this time. In both samples, meanwhile, the benefits of technology transfer remain contingent on prior link duration. This pattern may explain why supplier turnover has generally been higher in the U.S., but also why U.S. buyers have shown great loyalty to selected suppliers. A further difference pertains to the time that it takes for technology transfer to start paying off. It takes longer in the U.S. than in Japan (5.1 years versus 3.6 years). Our findings may also help explain differences in the propensity to rely on extensive technology transfer. Having experienced higher turnover in the decades leading to the 1990s, U.S. links are more recent on average but also consist of some very old ties (up to 92 years). The fact that many U.S. suppliers with shorter-lasting relationships expect little benefits from technology transfer plausibly helps account for the difference in technology transfer emphases between the two countries. Indeed, when both samples are split according to the median U.S. link duration, the difference between countries is less among cases with longer-established links.
In Japan, meanwhile, buyer-supplier links have started to fray in recent years. Relationships may now turn over faster, even for strong assemblers like Toyota. Thus, a challenge would again be to fortify recent links. While technology transfer starts paying off sooner, engaging in technology transfer too early is also problematic. Furthermore, our results show that small-scale technical exchanges do not improve performance relative to established Japanese competition. This implies a transition challenge when partners are replaced. The resulting tradeoff may explain the relative rigidity of Japanese buyer-supplier links. Unless relaxed, such rigidity could hinder adaptation in the face of global competition, radical technological change, or simply slower growth.
Building Relation-Specific Assets For U.S. and Japanese firms alike, then, a key challenge is how to build up relational assets to render technology transfer effective. We suggest two plausible approaches. One may be to focus on accelerating the idiosyncratic learning process whereby buyer and supplier develop joint understanding and routines. Adding feedback and making the information flow bilateral is a known way to accelerate learning through communication. The other solution is to leverage firms’ capacities to transfer technology, holding link duration constant. Prof. Martin and R. Salomon, a Stern alumnus now at the University of Southern California, have argued that two distinct capabilities contribute to successful interfirm knowledge transfer. Source transfer capacity pertains to a transferor’s ability to transmit knowledge outward, while recipient transfer capacity pertains to a transferee’s ability to assimilate knowledge from a willing external source. All things being equal, the most successful technology transfer will accrue in pairs of firms that possess the requisite combination of source and knowledge transfer capacity.
These findings shed light on the conditions for any convergence between U.S. and Japanese practice. Given the accumulated difference in mean link duration (10.6 years in our sample) and the residual difference in supplier turnover, adopting a straight “Japanese” model with extensive technological cooperation may be difficult and potentially wasteful for various U.S. firms. However, some U.S. firms are already in a strong position to leverage their technologies and relation-specific assets. If supplier relationships start turning over faster in Japan, meanwhile, U.S. practice might yield useful lessons – regarding technical exchanges in recent relationships, for example. Regardless of the context, two major lessons follow from our findings. First, suppliers stand to benefit from systematic knowledge exchange with buyers. Buyers, in turn, stand to benefit from a disciplined approach to knowledge exchange. Second, prior link duration conditions the effectiveness of more complex, higher-level technology transfer. Most important, higher-level technology transfer works best in long-established buyer-supplier relationships. The collaborative mechanisms we describe are not unique to vertical inter-firm relationships – or to the automotive industry. They also stand to affect performance inside firms that integrate vertically or diversify (especially via acquisition), in horizontal technology transfer deals between rivals, and in alliances. In each case, separate organizations must share knowledge for joint advantage to develop. This requires effective knowledge transfer mechanisms. Though governance and initial knowledge positions may vary, relation-specific assets stand to be critical in enabling the pooling of corporate capabilities. Masaaki Kotabe is the Washburn Chair of International Business and Marketing at Temple University; Xavier Martin, who conducted most of this research while on the faculty of NYU Stern, is associate professor of Organization and Strategy at Tilburg University (the Netherlands); Hiroshi Domoto is assistant professor at Tokyo University of Information Science. To find out more about the research reported here, see the authors’ extended paper, “Gaining from vertical partnerships: Knowledge transfer, relationship duration and supplier performance improvement in the U.S. and Japanese automotive industries,” in Strategic Management Journal, volume 24 issue 4, pages 293-316 (2003). |
n recent years,
researchers have devoted increasing attention to the effects of
supplier relationships on
buyers’ competitive advantage. By involving suppliers extensively
in product and process development, assemblers (buyers) can speed
product development cycles, lower input costs and boost end-product
quality. Able suppliers don’t simply manufacture parts according
to detailed specifications, they also help design the parts and the
corresponding manufacturing and technical processes. This division
of labor is accompanied by exchanges of knowledge about products
and processes to ensure suitable coordination. As a result, studies
have concluded that buyers should foster high-involvement relationships
with suppliers. 
eyond these main effects, research suggests
that buyers seek to further the benefits of knowledge transfer
by shaping the balance of technical
exchanges and technology transfer. Knowledge transfer is more difficult
when buyers and suppliers lack familiarity. Of course, we do not
expect the trend in a supplier’s performance, by itself, to
be inherently higher in longer links. Indeed, if parties to a longer-established
relationship develop a longer-term horizon shielding the link from
interruption in the presence of short-term performance dips, the
main effect of link duration may appear weak or slightly negative.
What’s more, newer relationships are able to exploit easy opportunities
for improvement while older partnerships must build past those by
tackling less obvious improvement projects. However, we expect longer-established
links, because they allow more relation-specific assets to develop,
to magnify some of the performance effects hypothesized above. That
leads to Hypothesis 3: The positive association between technical
exchanges and supplier performance improvement becomes stronger as
link duration increases.
ext we measured supplier performance by capturing
a suppliers’ performance
improvement relative to its position two to three years earlier,
in areas like product design, process design, product quality, and
lead-time. These performance dimensions are also consistent with
those used by assemblers to assess suppliers. These four items, when
introduced in the factor analysis alongside all the items that make
up our independent variables, loaded onto a single measure: Supplier
Performance Improvement. The measure of Technical Exchanges contained
six items pertaining to common, informal communication between engineers.
The measure of Technology Transfer consisted of five items describing
transfer of higher-level technological capabilities. The moderating
variable, Link Duration, was measured as the number of years since
the buyer and the supplier began their business relationship. 
or the Japanese companies, we found that the results support Hypothesis
4, and partially support Hypothesis 2, but do not support Hypotheses
1 and 3. Overall, technology transfer can be a powerful driver of
supplier performance, and its benefits accrue more in buyer-supplier
links that have been established longer. Further analysis provided
insight into the conditions for beneficial technology transfer. The
effect of Link Duration becomes positive only if it exceeds three
to four years. Interestingly, this is the typical length of a car
design and component purchasing cycle in Japan. Extrapolating the
result suggests the same challenge for new relationships as in the
U.S.: Relying on technology transfer too early may not be beneficial,
even though it is desirable in longer-established links. 
echnical exchanges may offer early benefits for U.S. firms, and should
thus be encouraged. However, given the contingent benefits of technology
transfer, a challenge is to make high-level technology transfer succeed
in recent relationships. Our results suggest that U.S. firms adopting
aggressive technology transfer practices prematurely are likely to
find these efforts comparatively ineffective. This makes it more
challenging to continuously upgrade technology.