Journal of Business Research 55 (2002) 33 – 40
The Japanese keiretsu system: an empirical analysis Jean McGuirea,*, Sandra Dowb a
Department of Management, Faculty of Commerce and Administration, Concordia University, 1455 de Maisonneuve West, Montreal, Quebec Canada H3G 1M8 b Department des Sciences Administratives, University of Quebec at Montreal, Montreal, Quebec, Canada Accepted 1 August 1999
Abstract What are the sources of ‘‘J form competitive advantage’’? This study uses ownership and financial networking variables to identify ‘‘J form’’ and ‘‘market-model’’ Japanese and US firms. Results using discriminate analysis indicate that ownership structure, inter-firm investment, and financing flexibility significantly differentiate J form firms from ‘‘market-model’’ firms. We then examine patterns in these market-model Japanese firms. We found that a higher proportion of keiretsu firms than non-keiretsu firms had adopted market-model ownership and financial networking characteristics. Further examination of these misclassified firms suggests that they may have responded differently to changes in the Japanese economy than correctly classified firms. D 2001 Elsevier Science Inc. All rights reserved. Keywords: Keiretsu; J form; Market model
1. Introduction The competitive strategy of Japanese firms is a topic of academic and popular interest. Sheard (1991, 1992, 1994) and Aoki (1990) suggest that the Japanese system of more flexible constraints on agency costs and blurred distinctions between stakeholders is an important source of Japanese competitive advantage. Specifically, they suggest that the Japanese keiretsu system of formal and informal links between firms, customers and suppliers, and financial institutions reduce agency costs and provide financial and strategic flexibility. These mechanisms differ from the ‘market model’ found in North America, which is characterised by market-based corporate control mechanisms and clear distinctions among financial stakeholders. Although often termed keiretsu organization, the widespread nature of inter-corporate networking in Japan suggests that both keiretsu and non-keiretsu Japanese firms benefit from inter-corporate networking (Caves and Uekasa, 1976; Price Waterhouse, 1989; Yamashita, 1989; Aoki, 1990; Prowse, 1990, 1992; Gerlach and Lincoln, 1992). We will refer to the ‘J form organization’ to describe this
* Corresponding author. Tel.: +1-514-848-2917; fax: +1-514-848-4208. E-mail address:
[email protected] (J. McGuire).
system. Two important characteristics of this system are stable shareholdings dominated by institutional and corporate shareholders and close ties with financial institutions. Institutional and corporate holdings dominate the ownership structure of Japanese firms. In 1990, financial institutions and corporate investors held 71% of outstanding shares in Japanese firms (Sheard, 1994). This compares to 51% individual ownership in the US (Aoki, 1990). By concentrating ownership in the hands of large financial institutions and corporations with multiple ties to the firm, the ownership structure of Japanese firms may promote efficient monitoring and control by equity holders without the short-term performance pressures typically found in North America (Sheard, 1994). Bank ties are another aspect of Japanese industrial organization. Japanese firms rely heavily on bank debt for financing needs (Corbett, 1994; Hoshi, 1994). Typically, Japanese firms will establish a relationship with a small number of banks, led by a ‘‘main bank’’ for a significant portion of their financing. Main bank relationships provide flexible financing and reduce the need to carry ‘reserves’ of long-term debt or liquidity (Brown et al., 1994; Corbett, 1994; Hoshi, 1994; Prevezer and Ricketts, 1994; Weinstein and Yafeh, 1998). Bank ties provide an additional source of monitoring through access to account records, contacts with firm management, and the like, not readily available to equity
0148-2963/02/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 1 4 8 - 2 9 6 3 ( 0 0 ) 0 0 1 2 8 - 4
34
J. McGuire, S. Dow / Journal of Business Research 55 (2002) 33–40
holders or other financial stakeholders. Although less often examined in the management literature, agency problems also arise between the firm and its creditors (Smith and Warner, 1979; Jensen and Smith, 1985; Stultz, 1990). Closer ties between the firm and its creditors, as well as the joint debt – equity positions often held by banks, allow banks and other creditors to better monitor the firm and reduce the agency costs of debt (James, 1987; Weinstein and Yafeh, 1998). In addition, a large percentage of Japanese firms are involved, to a greater or lesser extent, in established intercorporate networks. Both vertical and horizontal industrial groupings exist in Japan (Banjeri and Sambharya, 1996; Johnston and McAlevey, 1998). This study focuses on horizontal keiretsu, which has traditionally represented the corporate infrastructure of Japan. There are six commonly recognized horizontal keiretsu: Mitsubishi, Mitsui, Sumitomo, Fuji, Sanwa, and Dai-Ichi-Kangyo. Estimates place 89 of the 200 largest Japanese firms as having strong ties with one of several keiretsu groups (Hoshi et al., 1991). In fiscal 1992 corporations within the six families accounted for only 0.007% of all Japanese firms but almost 14% of total sales. (Japan Economic Newswire, July 26, 1994). In contrast to the access to supplies and innovation provided by exchange-based buyer – supplier networks, horizontal keiretsu provide more subtle advantages. Uzzi (1997) notes that economic action is embedded in social structure. He argues that socially embedded ties such as those found in horizontal keiretsu serve a more subtle regulatory and informational function. Congruent with these arguments, keiretsu ties may smooth and insulate the firm from short-term market pressures and the risks of capital market transactions, which provide the financial flexibility to focus on building long-term competitive advantage. Ownership is more closely intertwined for keiretsu firms, whose shares are held by ‘‘friendly insiders’’ and are infrequently traded on the open market (McCormack, 1994; Johnston and McAlevey, 1998). Keiretsu ties do not necessarily imply large individual stakes. Rather, keiretsu members as a whole dominate the ownership structure (Prowse, 1992; Sheard, 1994). These stable shareholdings insulate the firm from market pressures for short-term performance, corporate control activities, and the like, and provide ‘quasi inside’ information for member firms. These equity ties are complimented by other transactional links, which re-enforce the mutual interests of corporate stakeholders rendering short-term or self-serving behavior even more unlikely (Sheard, 1994; Gerlach & Lincoln, 1992). Banking relationships are higher among keiretsu firms. They also rely more heavily upon their ‘‘main bank’’ for borrowing than do non-keiretsu firms (Hoshi et al., 1991; Gerlach, 1992; Prowse, 1992). Beyond additional access to bank financing, bank ties allow keiretsu creditors to have more direct involvement in firm activities, and offer assistance in times of financial crisis (Morck and Nakamura, 1999; Kim and Limpaphayom, 1998).
Several authors, most notably Prowse (1990, 1992) and Williamson (1991a,b) propose the insulation and flexibility brought about by J form organization as a significant source of strategic advantage which could benefit North American firms. However, attempts by North American firms to adopt the J form organization face two obstacles. The more embedded historical and social ties found in Japan may be difficult to emulate in North America. More easily ‘imitatable’ ties such as equity links and closer links with financial institutions, isolated from this broader context, may not provide the same strategic benefits of more embedded ties (Uzzi, 1997). Others, however, have questioned the usefulness of the J form organization for North American firms. Johnson (1995) and Achrol (1997) suggest that the financial core of the keiretsu may be less appropriate in an increasingly market and information dominated strategic environment. J form organization may bring greater costs in the more competitive North American market. To illustrate these costs, Weinstein and Yafeh (1998) estimate that slightly over one-third of profitability differences between keiretsu and non-keiretsu firms can be explained by the higher costs of capital incurred by keiretsu firms when compared to unaffiliated counterparts.
2. Identifying J form firms in Japan and North America The preceding discussion suggests two important research questions: that of the ability of North American firms to emulate J form organization, and the continued strength of the Japanese inter-corporate network. To assess these issues, we operationalized J form organization in terms of ownership and networking characteristics. The Japanese context is characterised by numerous large ownership blocks. These large shareholdings are reinforced by smaller holdings by affiliated firms. Institutions are particularly prominent among these shareholders. Thus, we propose that J form organization will be reflected in both ownership concentration and institutional investment. ‘‘Network characteristics’’ refer to the less formal ties, which link keiretsu members without significant ownership stake in the firm. Examples of these network characteristics include: (1) historical links, informal and personnel ties; (2) closer, more flexible ties with sources of financing; and (3) inter-firm investment and other transactions. Williamson (1991a) and Sheard (1994) imply that such informal network ties may be of greater day-to-day importance than ownership stakes in promoting concern for long-term co-operation. Most obviously, the informal ties, and co-operation brought about by decades of collaboration may be difficult to emulate. Financial ties with institutions and other firms can also be an integral component of the networking strategy. Such mechanisms as inter-firm investment, sub-contracting and closer ties between suppliers, firms, and buyers are important elements of Japanese industrial organization (Sheard,
J. McGuire, S. Dow / Journal of Business Research 55 (2002) 33–40
1991; Williamson, 1994b). In North America, inter-firm investment, particularly in subsidiaries or suppliers, could be used to obtain keiretsu advantages. Such investment patterns have been suggested in the popular press as a means of facilitating co-ordination, co-operation and learning, and developing synergies among firms (Kelly and Port, 1992). Sheard (1994) further emphasizes the importance of reciprocal shareholdings and other transactional ties to the stability of the Japanese system. Close ties with sources of funding, particularly the ‘main bank’ is another central aspect of the Japanese inter-corporate network. Hoshi (1994) argues that these ties may have allowed Japanese firms greater access to external financing at lower cost and risk, making their investment patterns less subject to the availability of internal funding. We therefore suggest the following propositions regarding J form firms: Proposition 1 J form organization is characterised by greater ownership concentration of institutional investment than is ‘market-model’ organization. Proposition 2 J form organization is characterised by greater inter-corporate networking than ‘market-model’ organization. Proposition 3 J form organization is characterised by closer ties with banks and lending institutions than is market-model organization. The continued viability of ownership and networking mechanisms may be in doubt. The Japanese system may be evolving to a more market-model orientation due to the changes in product and financial market noted earlier (Corbett, 1994; Johnson, 1995). Increasing numbers of ‘market-model’ Japanese firms may indicate such evolution. It is uncertain whether keiretsu or non-keiretsu firms would be most subject to market-model pressures. Presumably less insulated non-keiretsu firms might be expected to be more open to such pressures. Alternatively, keiretsu firms may face greater pressure to adapt to the changing environment. We therefore advance a more general proposition: Proposition 4 Weakening of J form organization is reflected in Japanese firms exhibiting more market-model characteristics.
3. Research approach Our first three propositions identify ownership and financial networking as key dimensions of J form organization. To assess this proposition, we compared keiretsu firms, non-keiretsu firms, and US firms in terms of ownership and networking. Keiretsu firms would be expected to exhibit strongest J form characteristics, followed by non-
35
keiretsu firms. North American firms would be expected to exhibit ‘market-model’ characteristics. This analysis provided information as to the validity and relative importance of our J form variables (Propositions 1, 2, and 3). In other words, we would expect Japanese firms to exhibit greater networking, financial flexibility, and ownership concentration than US firms. To supplement this univariate analysis, the second phase of the research used discriminate analysis to classify firms into market model and J form categories based upon their ownership, networking, and financing flexibility profiles. As suggested by Proposition 4, however, not all firms in each national context would fit the hypothesised profiles. Some US firms would likely exhibit more J form characteristics, while certain Japanese firms may show greater market-model characteristics. This phase provided an additional test of the validity of our J form variables. More importantly, however, it allowed us to identify Japanese firms, which did not fit the expected J form profile. We then examined these firms in more detail to uncover possible reasons for their ‘‘atypical’’ ownership and network system. 3.1. Data and sample selection Sample firms were taken from the 1991 Disclosure and Worldscope databases. Worldscope is particularly useful in that it provides a standardised cross-national financial database (Brown et al., 1994). Additional data on US firms were taken from Disclosure for the same period. Single-year financial performance for 1991 was used in this analysis. Multi-year averages, when employed, cover the period 1988 to 1991. Foreign investment statistics for 1988– 1990 were obtained from the CFAR Database. 3.1.1. Japanese sample We used Nakatani’s (1984) classification of keiretsu and non-keiretsu (independent) firms, which has been used by previous researchers. Hoshi et al. (1991) and Gerlach (1992) have shown that keiretsu membership is stable and robust to classification criteria. Nakatani (1984) classified 248 firms as keiretsu firms and 69 as non-keiretsu. Our sample included 216 Japanese firms for which financial data were available of which 165 are keiretsu firms and 51 are nonkeiretsu firms. 3.1.2. North American sample The US sample consists of 610 firms. The sample was drawn from firms traded on the New York and American stock exchanges listed both in the Disclosure and Worldscope databases. The two samples were matched in terms of industry composition and size. 3.2. Operationalizing J form organization Operationalizing variables in cross-national research must consider the conceptual equivalence of measures, as
36
J. McGuire, S. Dow / Journal of Business Research 55 (2002) 33–40
well as their operational equivalence (Rosenzweig, 1994). A measure may be operationally equivalent (that is, measured consistently in two contexts), but may not be conceptually equivalent (have the same meaning or implications) in the two contexts. Within the parameters of this study, regulatory and reporting differences between the two countries also had to be considered. We therefore sought measures that (1) allowed for consistency of available data in both contexts; (2) tapped ownership or network dimensions relevant to keiretsu organization, and; (3) were comparable in both contexts. We therefore operationalized our J form variables as follows. 3.2.1. Ownership characteristics Ownership characteristics were measured by the percentage of shares held by significant blockholders and the percentage of shares held by institutional investors. A major characteristic of the Japanese system is more significant institutional ownership. Indeed, Williamson (1991a) and Porter (1992) place considerable emphasis on this ownership dimension. Institutional investors in both contexts are largely made up of mutual funds, insurance companies, and the like. One potential drawback of this measure may be that the role of institutional owners may differ between North America and Japan. There is mixed evidence to support this assertion (see, for example, Porter, 1992; Bathala et al., 1994; Corbett, 1994; Johnson, 1995). Balancing this are the numerous similarities in the role of institutional investment in Japan and the US. In both contexts, institutions are a significant, knowledgeable, and influential constituency. Ties between firms and institutional investors are stronger than ties with other investors (Baruch, 1992; Bathala et al., 1994; Corbett, 1994). Evidence suggests that institutional holdings are less frequently traded in both contexts and when traded, are more likely to remain in the hands of other institutional investors (Bathala et al., 1994; Corbett, 1994). We suggested earlier that the Japanese context allows input into firm decision-making at a lower ‘‘ownership threshold’’ than that found in the US due to numerous cross-holdings. Nevertheless, given more relaxed ownership restrictions and perhaps more limited market activity, we would anticipate that Japanese firms are more closely held than US firms even when raw measures such as the percentage of shares held by shareholders owning more than 5% of equity is employed. 3.2.2. Networking characteristics Networking characteristics supplement these more formal ownership ties. Many potential networking variables were not readily comparable between the two samples. For example, the ‘oversight’ function of the North American board of directors, reinforced by its legal mandate, differs from that of Japanese boards or the less formal ‘‘president’s council’’. Information on certain practices indicative of close ties among firms (for example, liberal trade credit,
informal extension or modification of the terms of credit) is not available. In view of these difficulties, network characteristics were defined in terms of inter-firm investment and flexibility in financing arising from closer ties with financial institutions. Ties to financial institutions are a key element of J form organization. Given the need to select a measure applicable and available to both sample groups we used the ratio of short-term debt to long-term debt as our measure of flexibility of financing. Several arguments support the use of this measure: 1. Empirically, Japanese firms make significantly higher use of short-term financing than do North American firms (Corbett, 1994; Hoshi, 1994; Prevezer and Ricketts, 1994). 2. Bank debt makes up a significant proportion of shortterm debt. Reliance on bank debt is a characteristic of the Japanese context. Moreover, banks have access to considerable ‘‘quasi inside’’ information about the firm’s finances (e.g., account records) on an ongoing basis, thus performing an important monitoring function (James, 1987; Goswami et al., 1994), given the low priority of bank debt during bankruptcy. The willingness of banks to continue to extend bank debt is seen as a signal of trust. 3. Short-term debt requires regular renegotiation. This is a double-edge sword. In market-driven situations, short-term debt usually carries a lower interest rate than long-term debt at the cost of renewal uncertainties. To the extent that such debt can be more easily rolled over in a co-operative (J form) context, both firms and banks get the advantage of flexibility with minimal additional cost. Thus, regardless of context, the use of short-term debt indicates the firm’s confidence in its ability to obtain future financing. Successful renegotiations of short-term debt may also foster greater on-going contacts between firms and financial institutions both in Japan and North America. Our second network variable, inter-firm investment was measured by the level of investment in subsidiary and affiliated firms calculated as a percentage of firm assets. This variable tapped the degree to which the firm has established equity ties with other firms. Blaine (1993) and Sheard (1994) emphasize that the network of reciprocal inter-firm investments is a critical element of the Japanese context and promote long-term co-operation Equity ties have also been cited as a means by which US firms can gain J form advantage (Kelly and Port, 1992). Two statistical techniques were used. Preliminary analysis was conducted using Analysis of Variance (ANOVA), which allowed us to examine whether the means of our four classificatory variables differed among the three sub-samples. Thus, we were able to determine whether keiretsu,
J. McGuire, S. Dow / Journal of Business Research 55 (2002) 33–40
non-keiretsu, and US firms differed in ownership, shortterm debt, and subsidiary investment. The second objective of the empirical analysis was to determine the prevalence of J form and market-model organization in both contexts. We used discriminate discriminate analysis to assess the extent to which Japanese and US firms conformed to their ‘expected’ multi-variate profiles. Discriminate analysis also allowed us to identify ‘typical’ and ‘atypical’ firms. We hypothesised that Japanese firms (in particular keiretsu members) which were empirically specified as market-model firms would be illustrative of an attenuation of J form cohesion.
4. Empirical results
37
Table 2 Discriminant analysis Predicted group membershipa Actual classification
Keiretsu
Non Keiretsu
US
Total
Count Keiretsu Non-keiretsu US
157 52 33
0 0 0
12 0 142
169 52 175
Percentages Keiretsu Non-keiretsu US
92.9 100 18.9
0 0 0
Test of functions
Wilks’ Lambda
c2
df
Sig.
1 through 2 2
0.42 0.99
342.64 5.21
8 3
0.000 0.16
7.1 0 81.1
100 100 100
Structure matrix
4.1. ANOVA results
Function
Results of the ANOVA analysis are presented in Table 1. For all variables (except for subsidiary investment, where US firms differed from only keiretsu firms) the mean of the US sample was significantly different from those of the two Japanese sub-samples. Contrary to expectations, US firms had higher ownership concentration in terms of both 5% owners and institutional ownership than either of our two Japanese sub-samples. These findings are congruent with Sheard’s argument that multiple ties make a significant ownership stake less necessary. As noted by Sheard (1994), other transactional ties can reinforce a smaller ownership position, allowing for a dispersed but concentrated ownership network. In essence, ownership stakes work more subtly through smaller but reciprocal ownership blocks re-enforced by multiple ties to suppliers, customers, and creditors. Although we had originally expected that Japanese firms would exhibit higher institu-
1
Institutional ownership Short-term debt Significant ownership Subsidiary investment
2 0.70 0.47 0.21 0.11
0.33 0.18 0.85 0.48
a
75.5% of original grouped cases correctly classified. Correctly classified counts and percentages are in bold italics.
tional investment, our results are consistent with ownership ties among J form firms being subtler in nature and therefore not contingent upon significant ownership stakes. Investment in subsidiary firms was higher in keiretsu firms. This higher level of inter-firm investment by Japanese firms likely reflects the more intense inter-firm links among keiretsu firms. As expected, Japanese firms make greater use of short-term debt than do North American
Table 1 Comparison of means
Keiretsu
Non-keiretsu
US
Total
Institutional investment*
Significant ownership**
Subsidiary investment***
Short-term debt****
M = 27.37 (8.42) N = 169 M = 24.34 (9.77) N = 54 M = 48.78 (18.94) N = 182 M = 36.58 (18.03) N = 405
M = 20.26 (15.99) N = 169 M = 13.29 (14.38) N = 54 M = 27.3776 (24.4074) N = 182 M = 22.53 (20.60) N = 405
M = 3.26 (5.70) N = 171 M = 2.65 (3.52) N = 55 M = 1.71 (4.56) N = 178 M = 2.49 (5.01) N = 404
M = 47.77 (22.57) N = 170 M = 51.739 (27.03) N = 53 M = 23.64 (22.17) N = 77 M = 37.62 (26.17) N = 400
For all variables except subsidiary investment, US group means are significantly different from both Japanese samples using Tukey post-hoc comparison tests. For subsidiary investment only the US – Keiertsu comparison was significant. * F statistic 121.57 significant at 0.01 level. ** F statistic 12.12, significant at 0.01 level. *** F statistic 4.32, significant at 0.02 level. **** F statistic 59.06, significant at 0.01 level.
38
J. McGuire, S. Dow / Journal of Business Research 55 (2002) 33–40
Table 3 Comparison of correctly classified and market-model Keiretsu firms
Correctly classified Market-model keiretsu Total
5-year return on assets
Sales growth
Employee growth
Foreign ownership*
3.91 (1.61) N = 157 4.28 (1.86) N = 12 3.94 (1.62) N = 169
8.41 (7.16) N = 157 6.90 (7.52) N = 12 4.03 (5.37) N = 169
1.35 (6.21) N = 157 2.34 (2.89) N = 12 1.42 (6.03) N = 169
3.81 (5.13) N = 154 6.90 (7.52) N = 12 4.03 (5.37) N = 166
* F ratio 3.76 significant at 0.05.
firms. Differences between the two Japanese sub-samples were insignificant. These results tend to support hypothesised differences between Japanese and North American firms in our classificatory variables. That the means of the two Japanese subsamples are not significantly different suggests that both types of Japanese firms manifest the more co-operative norms of the Japanese context. We might have expected
that non-keiretsu firms would represent a ‘middle ground’ between keiretsu and US firms. The pattern, however, is interesting and contrary to expectations: for all variables except short-term debt the mean for keiretsu firms fell between that of US firms and non-keiretsu firms. 4.2. Discriminate analysis results ANOVA examines only differences in sub-sample means on each variable. It does not examine whether a combination of the three variables allows us to distinguish among our three sub-samples. Discriminate analysis evaluates such multi-dimensional profiles. Results of the discriminate analysis are presented in Table 2. The model correctly classified 75.5% of the overall sample. Correct classification was highest among keiretsu firms, with 92.9% correctly classified, followed by US firms (81.1%). However, none of the non-keiretsu firms were correctly classified. This pattern suggests that although the model is applicable to both the Japanese and North American contexts, it is difficult to discriminate between the two classes of Japanese firms using our classificatory variables.
Table 4 Paired comparisons of market-model keiretsu firms Misclassified firm
News summary
Comparison firm(s)
News summary
Dai Nippon Pharmaceutical
Biotechnology firm showing 17.4 profit decline for 1991; eight cooperative ventures during 1991 – 1992 period Sales and profit decline in second quarter, 1992 FYb 1991 profit decrease of 37%
Banyu Pharmaceutical Nippon Kayaku
Significant (25 – 30%) profit declines in both firms, three (Banyu) and four (Nippon) co-operative ventures 6.4% profit increase for 6 months November, 1992 81% decline in profits 1992, 49% drop in 1991 39.5% decrease in net profits for fiscal year 1992
Dai-ichi Kogo Seiyku Kokusai Electric Matsuo Bridge
Sanyo Chemical Alps Electric
Kayaba Industries
FY 1991 Net profit decline of 65.9%; pretax profits down 57.1%
Toto
Profit increase of 11% for 1991 March 1992 ‘off floor’ trading of shares to increase number of shareholders FY 1991 Operating profit fell 36.9% yen, current profit dropped by 24.6%; Net profit also decreased by 11.2%; FY 1990 operating profits increased by 5% 2% No reports on firm
Noritakea
Toyo Ink
Slight profit decline
Dainippon Inka
Yamatake Honeywell
First half 1991 profit decline of 14%; expected decline of 36% for fiscal 1991
Shimadzu
31% drop in nonconsolidated 6-month profit for early 1992; 5.4% drop in profits for April – September, 1991 Pre-tax profit decline of 34% for half year ending September, 1992; June 1992 yearly net profit decline of 54% 32% profit decline for 1991
Nippondenso
Miyaji Iron Works
a Since not all misclassified firms could be matched with sample firms, comparisons for these firms were taken from Worldscope. Noritake was the closest ‘match’ in Worldscope for Toto, while Dianippon Ink was the closest match for Toyo Ink. b FY=fiscal year.
J. McGuire, S. Dow / Journal of Business Research 55 (2002) 33–40
These findings provide only limited support for Proposition 4 in that only a small percentage of Japanese firms were misclassified as being US firms. Although non-keiretsu classification might have been expected to represent a ‘mid-point’ between keiretsu and US characteristics, this does not appear to be the case. It is significant to note, however, that misclassified keiretsu firms were placed in the US catetory, whereas misclassified non-keiretsu firms were classified as keiretsu members. This pattern provides some support for the argument that keiretsu firms have been more likely to diverge from J form ownership and networking characteristics.
5. ‘‘Market-model’’ keiretsu firms We now focus on keiretsu firms incorrectly classified as US. The purpose of this analysis was to discover patterns that might explain their adoption of market-model characteristics. We first examined patterns in keiretsu affiliation among misclassified firms. Representation of each keiretsu was similar, with no industry trends apparent. Another possible explanation of misclassification might be exposure to ‘market-model’ pressures due to foreign ownership or poor performance. We compared levels of accounting performance, employee growth, and foreign ownership as shown in Table 3. Market-model keiretsu firms had significantly higher levels of foreign ownership. These findings may suggest that foreign ownership creates pressures for market-model characteristics. Alternatively, foreign owners may be attracted to market-model firms. Regardless, these findings suggest higher foreign ownership influence among misclassified firms (Table 4). The statistical results provide little evidence of performance differences between market model and correctly classified firms. With only 12 market-model firms, however, we continued with a more qualitative analysis. We searched the Japanese Economic Newswire, which reports performance figures, earnings and performance estimates, strategic initiatives, and the like. The time period January 1, 1991 to December 31, 1992 was selected to provide an adequate ‘window’ to identify possible explanatory factors. We matched each market-model keiretsu firm with a correctly classified firm operating in the same primary SIC code. No matching sample firm was available for two of the misclassified firms. For these firms, comparison firms matched by SIC firms were taken from Disclosure. With the possible exception of Dai Nippon Pharmaceutical’s high level of co-operative ventures, there appears to be no systematic difference in the strategic initiatives undertaken by market-model firms. Interestingly, however, reports of profitability suggest a trend toward higher reported and estimated profitability among market-model firms. With the exception of Dai-Ichi Kogo, financial reports by marketmodel firms seem to be more favorable than those of comparison firms operating in the same industry. While
39
no differences were found in ROA and five-year growth trends, news reports of the performance of market-model keiretsu firms appear to be more favorable than for matched firms in the same industry. This points to the possibility that performance differences may be associated with the adoption of market-model organization.
6. Conclusions Our analysis suggests several conclusions. First, our network and ownership variables were able to successfully classify a large proportion of our sample. Large blockholdings and institutional investment favored market-model classification. Given the strong theoretical argument for the role of institutional investment in the Japanese context, analysis of different classes of institutional investors may provide additional insights into the role of institutional investment in Japan. This finding, however, provides little support for arguments that liberalized equity participation on the part of financial institutions may facilitate ‘Japanesetype’ industrial organization. Indeed, an opposite trend may occur. Increases in the already higher levels of institutional and large block ownership may only aggravate existing stakeholder pressures. Our network variables were also able to differentiate Japanese and North American firms. In view of the preliminary nature of our networking variables, these results suggest the need to further develop understanding of the more informal aspects of inter-corporate networking. Future research should attempt to incorporate the less formal aspects of networking. Not only do these ties play a critical role in Japanese intercorporate networks, they may also be less subject to external and internal pressures than more instrumental exchange relationships. Our analysis of market-model Japanese firms suggested several possible explanations for their more market-model characteristics. The higher levels of foreign ownership found among marketmodel keiretsu firms is suggestive of a gradual opening of the Japanese industrial context which may, eventually, impact on informal networking. Our measurement of J form mechanisms was preliminary in nature and focused on financial networking. It is difficult to assess the extent to which misclassification stemmed from specification errors in the model, or to changes in Japanese industrial organization. Such changes in ownership may be precursors of further changes in corporate strategies. Implicit in stakeholder theories of the firm is that the preferences of important stakeholders’ influence firm actions. If this is indeed the case, the Japanese corporate context may increasingly reflect the preferences of these stakeholders. Our measures did not tap the range of network ties possible in both the US and Japanese contexts. As noted in our discussion, our research exhibits many of the common limitations of cross-national studies in that our measures had
40
J. McGuire, S. Dow / Journal of Business Research 55 (2002) 33–40
to balance theoretical relevance and practical exigencies. More refined measures might have more accurately tapped the networking activities of these firms. However, our findings support the theoretical relevance of the networking and ownership dimensions of J form organization. Finally, we did not address the strategic and performance implications of the J form organization The impact of J form organization on firm strategy is an important area of future research. For example, Ito (1995) and Banjeri and Sambharya (1996) provide evidence that keiretsu membership may influence market entry strategies and firm –subsidiary relationships. The strategic implications of J form organization in both Japan and the US are an important area of future research. This research would address the extent to which these strategic implications of J form organization are context-specific.
Acknowledgments An earlier draft of this paper was presented at the 1996 Academy of International Businesses Meeting in Banff, Alberta. This manuscript benefited greatly from comments by the editor and reviewers. This research was supported by a grant from the Social Sciences and Humanities Research Council of Canada.
References Achrol R. Changes in the theory of inter-organizational relations in marketing: toward a network paradigm. J Acad Mark Sci 1997;25:56 – 71. Aoki M. Toward an economic model of the Japanese firm. J Econ Lit 1990;8(March):1 – 27. Banjeri K, Sambharya R. Vertical keiretsu and international market entry: the case of the Japanese automobile ancillary industry. J Int Bus Stud 1996;27(First Quarter):89. Baruch L. Information and disclosure strategy. Calif Manage Rev 1992;34(Winter):9. Bathala C, Moon K, Rao R. Managerial ownership, debt policy and the impact of institutional holdings. Financ Manage. 1994:38. Blaine M. Profitability and competitiveness: lessons from Japanese and American firms in the 1980s. Calif Manage Rev 1993;36(Fall):48 – 65. Brown PR, Soybel VE, Stickney CP. Comparing US and Japanese corporate level operating performance using financial statement data. Strategic Manage J 1994;15(January):75 – 83. Caves R, Uekasa M. Industrial organization in Japan. Washington, DC: The Brookings Institute, 1976. Corbett J. An overview of the Japanese financial system. In: Dimsdale N, Prevezer M, editors. Capital markets and corporate governance. Oxford: Oxford Univ. Press, 1994. pp. 306 – 24. Gerlach M. The Japanese corporate network: a block model approach. Administrative Sci Q 1992;37(January):105 – 39. Gerlach M, Lincoln R. The organization of business networks in the U.S. and Japan. In: Nohria N, Eccles R, editors. Networks and organization. Cambridge, MA: Harvard Business School, 1992. pp. 451 – 520. Goswami G, Noe T, Rebello M. Debt financing and myopia: cause or consequence. Working paper 94-12, Department of Finance, College of Business, Georgia State University, University Plaza, GA. 1994.
Hoshi T. The economic role of corporate grouping and the main bank system. In: Aoki M, Dore M, editors. The Japanese firm: sources of competitive strength. Oxford: Oxford Univ. Press, 1994. pp. 285 – 309. Hoshi T, Kashyap A, Scharfstein D. Corporate structure, liquidity and investment: evidence from Japanese industrial groups. Q J Econ 1991;106(February):33 – 59. Ito K. Japanese spin-offs: unexpected survival strategies. Strategic Manage J 1995;16(September):431 – 47. James C. Some evidence on the uniqueness of bank loans. J Financ Econ 1987;19(December):217 – 35. Japan Economic Newswire, Various Editions. Jensen M, Smith C. Stockholder, manager, and creditor interests: applications of agency theory. In: Altman E, Subrahmanyam M, editors. Recent advances in corporate finance. New York, NY: Irwin, 1985. Johnson S. Managerial dominance of Japan’s major corporations. J Manage 1995;21:191. Johnston S, McAlevey L. Stable shareholdings and Japan’s bubble economy: an historical overview. Strategic Manage J 1998;19(November):1101 – 7. Kelly K, Port O. Learning from Japan. Bus Week 1992;27(January): 52 – 8. McCormack J. The Japanese way: the relationship between financial institutions and non-financial firms. Government of Canada, Economic, and Trade Policy, Policy Staff Paper-No. 94/16. (June 1994). Morck R, Nakamura M. Banks and corporate control in Japan. J Finance 1999;54(February):319 – 39. Nakatani O. The economic role of financial corporate grouping. In: Aoki M, editor. The economic analysis of Japanese firms. Amsterdam: NorthHolland, 1984. Porter ME. Capital disadvantage: America’s failing capital investment system. Harv Bus Rev 1992;70(Sept. – Oct.):65 – 82. Prevezer M, Ricketts M. Corporate governance: the U.K. compared with Germany and Japan. In: Dimsdale N, Prevezer M, editors. Capital markets and corporate governance. Oxford: Oxford Univ. Press, 1994. pp. 237 – 56. Price Waterhouse. Doing business in Japan, New York, 1989. Prowse SD. Institutional investment patterns and corporate financial behaviour in the United States and Japan. J Financ Econ 1990; 27(July):43 – 66. Prowse SD. The structure of corporate ownership in Japan. J Finance. 1992;47(3):1121 – 40. Rosenzweig P. When can management science research be generalized internationally? Manage Sci. 1994;40(1):28 – 39. Sheard P. The economics of Japanese corporate organization and the ‘‘structural impediments’’ debate: a critical review. Jpn Econ Stud 1991;19:30 – 78. Sheard P. International adjustment and the Japanese firm. Sydney, Australia: Allen and Irwin, 1992. Sheard P. Interlocking shareholdings and corporate governance. In: Aoki M, Dore M, editors. The Japanese firm: the sources of competitive strength. Oxford: University Press, 1994. pp. 310 – 49. Smith C, Warner J. On financial contracting—an analysis of bond covenants. J Financ Econ 1979;7(June):117 – 61. Stultz R. Managerial discretion and optimal financing policies. J Financ Econ 1990;26(July):263 – 92. Uzzi B. Social structure and competition in interfirm networks: the paradox of embeddedness. Administrative Sci Q 1997;42(March):35 – 67. Weinstein D, Yafeh V. On the costs of a bank-centered financial system: evidence from the changing main bank relations in Japan. J Finance 1998;53(April):635 – 72. Williamson OE. Comparative economic organization: the analysis of discrete structural alternatives. Administrative Sci Q 1991;36(June):269 – 96. Williamson OE. Strategizing, economizing, and economic organization [special issue]. Strategic Manage J 1991;12:75 – 94. Yamashita T. Japan’s securities markets. Hong Kong: Butterworth, 1989.