
































Microsoft Word - APE-V4N2-p44


Advances in Politics and Economics 
ISSN 2576-1382 (Print) ISSN 2576-1390 (Online) 

Vol. 4, No. 2, 2021 
www.scholink.org/ojs/index.php/ape 

44 
 

Original Paper 

Japanese Network Economy-ERGM analysis of the “Big Six” 

Marton Gosztonyi1* 
1 Budapest LAB Enterprise Development Center, Budapest Business School University of Applied 

Sciences, Budapest, Hungary 

* Márton Gosztonyi, Research fellow, Budapest LAB Enterprise Development Center, Budapest 

Business School University of Applied Sciences, Budapest, Hungary 

 

Received: May 15, 2021         Accepted: May 23, 2021        Online Published: May 24, 2021 

doi:10.22158/ape.v4n2p44            URL: http://dx.doi.org/10.22158/ape.v4n2p44 

 

Abstract 

The paper analyses the network of relationships between large companies that defines the Japanese 

economy using ERGM (Exponential Random Graph Model) method. The Japanese economy, because 

of the presence of keiretsu groups and other corporate groups, is an excellent example of what can 

happen to a country’s economy if large corporations and financial institutions in the market operate in 

a closely interconnected, highly centralized and dependent network. This is an important issue, because 

this economic practice is happening globally right now. Therefore, the interpretation of the Japanese 

network and the analysis of its economic performance also reflect the long-term negative and positive 

effects of this trend. 

Keywords 

Japanese economy, network economy, ERGM 

 

1. Introduction 

As the economy and economic actions are always embedded in society (Polányi, 1944, 2001), the 

functioning of different national economies, in addition to the global context, is always determined by 

the history, social relations and social network of the given culture. The economies of nation-states in 

the twenty-first century belong to the age of capitalist economies (Note 2), where the corporate and 

financial sectors are becoming more and more interdependent. Based on all this, it is worth looking at 

an economy like that of Japan’s, operating in a close and dense network, (in which the engines of the 

economy are still based on a dense network of large corporations and banks), and analyzing the 

network factors that explain their operation. 

The literature points to the collapse of Japan’s bubble economy in the 1990s, followed by the long 

economic stagnation over the past 30 years. All of this was only exacerbated by the 2008 global 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

45 
Published by SCHOLINK INC. 

economic crisis, the catastrophic earthquake of 2011, and the tsunami that damaged northeastern Japan 

(Yoshino et al., 2014). Despite all this, the Japanese economy, of course, continues to be a leader 

among world economies, ranking third in the world in terms of GDP per capita (World Bank, 2020). 

However, several studies point out that Japan’s economic miracle and stagnation have the same 

components, viz., a strong bond between the corporate and banking sectors. The Japanese banking 

system is characterized by different types of banks: long-term credit banks, asset management banks, 

city banks, regional banks, secondary regional banks, and insurance companies (De Masi et al., 2011). 

These organizations and banks have a key role to play in the Japanese network economy. In each case, 

Japan’s “big six” (Note 3) have an “internal bank” that is tied to the corporate group with multiple 

threads. Not only do these central financial institutions own more than 50% of the shares in firms 

(Agarwal & Elston, 2001), but in many cases, the chairs of supervisory boards are also bankers or 

bank-related actors who are part of corporate management. All this results in a classic merger of 

banking and market companies, analogous to the classical German banking system of the 19th century. 

Sterken and Tokutsu (2002) have shown that this connection encourages conglomerates to take out 

more favorable loans, and that the presence of a credit in itself attracts more loans from other banks. All 

this also results a number of phenomena that strengthen and distort the markets. There are long-term 

and close credit relationships between banks and large corporations, resulting in a highly hierarchical 

structure in which the central network positions are occupied by leading banks. However, this is 

accompanied by a strong asymmetric information problem in the market (De Masi et al., 2011). As 

Japanese firms operate in this system, they are heavily dependent on banks’ funding, leading to the 

bankshaving much more information about the firms than their counterparts in other countries’ credit 

markets (Spiegel & Yamori, 2003). In contrast, banks provide preferential credit terms, which are more 

lenient in terms of easier access to capital, bank debt settlement and the possibility of other preferential 

loan terms (De Masi et al., 2011). It follows from all this that in Japan, banks are steering firms towards 

low-risk investment decisions, which also implies that firms closely related to banks are less profitable 

than their other competitors (Ogawa et al., 2007). Ogawa et al. (2007) provided a detailed analysis of 

the evolution of the number of long-term debt ratios and the relationship between key indicators of the 

firms; the authors highlighted that although the largest firms have the most banking links, the number 

of links is positively correlated with solvency and R&D, but has a negative effect on the firms’ 

liquidity. 

Japanese financial institutions suffered heavy damage after the 1990 Bubble collapse due to the 

above-mentioned structural constraints. There has been a significant deterioration in the financial status 

of the Japanese banking system since 1992 (Brewer et al., 2003), with commercial banks recording a 

total loan loss of about Yen 83 trillion. These losses have reduced the capitalization of banks and led to 

the failure of three large banks. The precarious financial situation of these large banks affected the 

entire credit system, especially their weaker financial counterparts. In order to increase the financial 

stability of the system, Japanese regulators liquidated a city bank and nationalized two of the three 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

46 
Published by SCHOLINK INC. 

largest long-term credit banks (Brewer et al., 2003) in 1997. The failures of bank negatively affected 

the stock prices of the credit-related businesses as well (De Masi et al., 2011). 

Therefore, a major financial transformation was implemented in Japan in the 2000s. The CRL 

(Commercial Rehabilitation Law) was enacted, which restructured the financing system’s 

network-based connections, which lead to positive results by 2012. This was followed by further 

economic stimulus decisions by the government, decentralizing the system and removing the ties of 

market players. All this, in financial terms, has had positive results, as stock market prices have started 

to rise after a long stagnation (Chan-Lau, 2002). Of course, this result also required Japan to increase 

the share of foreign capital investment, compared to the national capital investments that had 

dominated the markets so far. All this has been recorded in literature, with many authors studying the 

FDI and its results like increased employment, output, and productivity (Driffield, 2001; Driffield & 

Girma, 2003; Driffield & Munday, 2000; Driffield & Taylor, 2000; Kimino et al., 2007). 

Japan has taken a huge step towards tackling growth-restrictive corporate practices such as closed-loop 

sales, cartel-like behavior and entrenched agreements, which makes it more difficult for new entrants to 

compete for products, services and / or prices (Yoshino et al., 2014). 

 

2. Zaibatsu and Keiretsu 

It is important to note, however, that the afore said image captures only a part of the story. The 

development of this level of network hierarchy in the banking system is the result of a long social 

change leading all the way back to the operation of Edo-era family businesses. However, I would not 

interpret the whole economic history of Japan in this paper, but confine myself to the development of 

the capitalist economy. 

Perhaps one of the most important internal engines of the development of Japanese capitalism (Showa 

era (1926-) was the Zaibatsu. Japan arrived in the global capitalist economic scenario relatively late, 

and even then, the process began due to external pressure (Imai, 1987). The engines of this economic 

transformation were the Zaibatsus, created to supplement the market imperfections of the changing 

economy. As a result, Zaibatsu firms, which have optimized their mutual returns within a strong 

hierarchical system, can be interpreted as a form of early-capitalist Japanese innovation through closely 

inter-related activities (Imai, 1987). These economic networks have created a strong control structure 

through asset-share ownership, human relationships, trust, and centralized purchase and marketing. The 

basic method of control was direct share ownership (Imai, 1987). 

The close relationship between firms was reflected in the fact that executives held leadership roles in 

more than one company simultaneously, such that these organizations could “function as a monolithic 

whole” (Hadley, 2015). As a result, Zaibatsu has been able to bring success to several industries in the 

Japanese economy. This is how the Zaibatsu of 294 companies operating under Mitsui, the Zaibatsu of 

241 companies under Mitsubishi, and that of 166 companies under Sumitomo emerged (Imai, 1987). 

This corporate transformation coincided with the transfer of technology from developed economies, 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

47 
Published by SCHOLINK INC. 

especially from Britain, to less developed economies, including Japan. The leaders of the successful 

Zaibatsu also had a missionary zeal in developing Japan as an industrial nation. Therefore, they 

opposed short-term gains and emphasized risk aversion, making strategic investments for 

industrialization (Imai, 1987). By the first half of the twentieth century, Zaibatsus became key players 

in bank subsidies, creating a close link and entanglement between the financial and corporate sectors 

(De Masi et al., 2011). 

However, in World War II, these groups of companies opted for a strong political role, resulting in the 

economic cleansing announced in January 1947 after the capitulation of Japan. As is well known, 

senior executives of 245 large Japanese companies were removed. This was an epoch-making event, as 

it disrupted the Zaibatsus’ networks-based transactions on personal attachment and changed the way 

decisions were made within the network groups. As a result, people in their forties who were at the 

level of department heads were appointed to the highest levels of governance, and they began to invest 

in innovative segments of the Japanese, economy, putting it on a turbulent trajectory. After the war, 

Japan thus became one of the fastest growing economies in the industrialized world. 

This process also went hand- in- hand with the process of Keiretsu groups taking over the role of 

Zaibatsus. Yonekura’s (1985) corporate graph typology (Figure 1) perfectly illustrates the differences 

between the two structures. Keiretsu groups show a more decentralized and reciprocal relationship 

structure, with a strong core that forms a reciprocal and decentralized network and a periphery made up 

of companies belonging to one of the core actors and with which they have a hierarchical and one-way 

relationship. 

 

 

Figure 1. Schematic Graphs of the Zaibatsu and Keiretsu Groups 

S. Yonekura (1985, p. 64). 

 

One of the special features of the rapid growth of the Japanese corporate system was therefore the 

Keiretsu groups. Derived from the previous historical narrative, they cover a network of close and 

hierarchical corporate financial groups, in which banks have developed a close relationship of 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

48 
Published by SCHOLINK INC. 

dependence with companies. In many cases, the group's main bank also played the leading role in 

Keiretsu. There is also a stable internal cross-shareholding between the players in the Keiretsu group, 

similar to the Zaibatsu, with each company owning shares in the other companies in the group. 

Companies in Keiretsu usually do not sell their shares to outsiders. Additionally, the central bank 

usually owns the largest equity fund and is also the main creditor to the other members of the group 

(Chan-Lau, 2002). 

Keiretsu groups are classified in Japan as the Big Six, viz., Sumitomo, Mitshubishi, Fuji, Sanwa, Mitsui 

and Daiichi Kangyo. Figure 2 shows the current group of big six and the most important companies 

belonging to them. 

 

 

Figure 2. The Big Six and Their Leading Companies 

Lincoln, 2004, 22. 

 

This corporate governance system was one of the most important contributors to the success of the 

Japanese economy from the early 1950s to the late 1970s. However, downturns appeared in the central 

banking system by the mid-1970s, and the slow liberalization and regularization of the financial system 

began. Large and creditworthy companies began to reduce the use of bank loans as a major source of 

financing, as they were able to obtain cheaper finance in the securities markets. The concentration of 

bank loan portfolios began to shift towards medium-sized and smaller firms, which were in many cases 

independent of Keiretsu groups (Chan-Lau, 2002). 

By the 1980s, although the Japanese economy was still on a turbulent trajectory, these downturns 

become stronger. An increasing problem which emerged by then was the corporate boards being were 

aligned better with creditors or banks than with employees or shareholders (Kanda, 1998), making 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

49 
Published by SCHOLINK INC. 

them more interested in strong bank support than in profit maximization (Morck &Nakamura, 1999). 

Further, other economic reforms also began in the 1980s, such as reducing the number of company 

executives and increasing focus on their performance, rather than on the length of their tenure in 

management positions., but this still wasn’t enough that the dotcom burst don’t drag down the Japanese 

economy with itself by the early 1990s, followed by a stagnant trajectory ever since (Chan-Lau, 2002). 

  

2. Method 

2.1 ERGM Network Analysis 

Due to the strongly networked economic structure outlined above, a number of network studies have 

already been conducted about the Japanese economy, both descriptively and in a bipartite fashion 

(Peltomaki-Alava, 2006; Sneppen et al., 2004; Guillaume-Latapy, 2004). The main network analysis 

directions of the Japanese economy are the analysis of corporate debt and lending processes and 

network clusters (De Masi et al., 2011), in cross-sectional and longitudinal ways. 

In my paper, however, I examine the Keiretsu groups based on ERGM (Exponential Random Graph 

Models) network analysis, a technique that can be classified as statistics- based network model building. 

ERGM allows the joint and controlled study of the effects of network parameters. In other words, 

ERGM models provide an opportunity to understand how and why network connections are formed. 

Network connections are organized into patterns, and the presence of some connections (ties) promotes 

the appearance of others. The ERGM analyzes these endogenous effects, often referred to as 

“structural” effects. Thus, it interprets the internal processes of the network relationship system, 

complemented by the exogenous effects of the network, such as the effects of attributes associated with 

nodes (Lusher et al., 2013). Based on the findings of Watts (1999), ERGM modeling works with 

randomness and probabilistic random graph modeling. By incorporating randomness, ERGM models 

work with expected values, from which we can draw conclusions about whether the observed data are 

consistent with the theoretical (anticipated) data. During an ERGM modeling, we estimate the presence 

of network edges from several predictor variables and use model parameters to estimate their effect, 

direction, and significance in the studied network (Lusher et al., 2013). 

For this reason, ERGM tries to model the network effects of edges, e.g. effect of transitive triads, k-star 

effect, effect of assertively (hemophilic, heterothallic relationships), effect of distributions, effect of 

degrees and effect of attribute variables. Further, it compares the models formed on the basis of the 

effects with the observed network and thus tries to reveal the causes of network edge formation. 

In this paper, network descriptive statistical and morphological analyses were performed with Gephi 

0.9.2v software, and ERGM analyses were created with the R software package and MPNet software 

(Wang et al., 2009). 

2.2 Data used for Analysis 

The data used for the ERGM model rely on a secondary data analysis from Lincoln et al. (2004). In 

their 1992 and 2004 works, Lincoln and co-authors performed network-based block modeling of 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

50 
Published by SCHOLINK INC. 

Japanese keiretsu groups, where networks were grouped based on the attributes of connections and 

nodes. In their longitudinal study, clusters of keiretsu groups from 1978 to 1997 were analyzed across 

the entire leading corporate population (258 firms). They used several variables to define the formation 

of the edges (ties), such as partial ownership, cross lending, capital, distribution of director positions. 

Then based on these, they formed the blocks using an algorithm called CONCOR (Note 1). Thus, 

clusters were created that contained closely or densely connected nodes (corporations). 

I used the data of the 1997 keiretsu block model in my paper, and built the economic network based on 

this data. In the network, the companies are classified into six large Keiretsu groups and into two large 

economic blocks. One of the large economic blocks is the banking and finance block and the second is 

the industrial block. In addition, nodes also have an attribute indicating their economic sector. 

If we consider the blocks, the financial block can be further divided into 5 clusters and the industrial 

block into 14 clusters. Based on the distributions in Table 1, it can be clearly seen that the companies 

belong to the industrial cluster have an overwhelming majority in the distribution of companies 

(81.01%), with a much smaller (18.99%) financial cluster 

 

Table 1. Block Model of Keiretsu Groups (N = 258) 

Blocks and Clusters 

Block and Cluster 

Frequency % 

Financial Block 49 18.99 

City and Trust Banks 19 7.36 

Mining 1 0.39 

Mutual BKS and Securities 9 3.49 

Regional banks and Insurance 16 6.20 

Toyota 4 1.55 

Industrial Block 209 81.01 

DKB And Sanwa Group 17 6.59 

DKB Group 15 5.81 

Fuyo and Sanwa groups 15 5.81 

Fuyo Group 16 6.20 

Hitachi 7 2.71 

Matsushita Group 13 5.04 

Mitsubishi and DKB 14 5.43 

Mitsubishi Group 18 6.98 

Mitsui Group 33 12.79 

Sanwa and DBK Group 12 4.65 

Sanwa Group 13 5.04 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

51 
Published by SCHOLINK INC. 

Sumitomo Group 13 5.04 

Toyota Group 11 4.26 

Trading Companies 12 4.65 

Grand Total 258 100.00 

 

In order to interpret the ERGM data, it is also worth reviewing the distribution of nodes by Keiretsu 

groups (Table 2). The Table shows that the 6 large Keiretsu groups comprise about the same number of 

leading companies (6-10 percent of the total companies). It is also important to note that the share of 

independent companies (not part of any Keiretsu group) is quite high, at more than 55% of the total 

market. 

 

Table 2. Distribution of Keiretsu Groups (N = 258) 

Groups Frequency % 

DKB 25 9.69 

Fuyo 19 7.36 

Independent(shacho-kai) 147 56.98 

Mitsubishi 16 6.20 

Mitsui 16 6.20 

Sanwa 24 9.30 

Sumitomo 11 4.26 

Total 258 100.00 

 

In addition to the above network edges, one more attribute parameter was included in the ERGM model, 

which indicated the economic sector of the companies. The distributions of this variable are 

summarized in Table 3. 

 

Table 3. Distribution of Cooperation by Economic Sector (N = 258) 

Sectors Frequency % 

Automobile 17 6.59 

Banking 49 18.99 

Ceramics 8 3.10 

Chemicals 28 10.85 

Electronics 22 8.53 

Fishing 4 1.55 

Food 26 10.08 

Forestry  1 0.39 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

52 
Published by SCHOLINK INC. 

Gen manuf 5 1.94 

Heavy metal 15 5.81 

Light metal  14 5.43 

Machinery 13 5.04 

Mining 2 0.78 

Oil 8 3.10 

Paper 7 2.71 

Pharma 5 1.94 

Precequip 2 0.78 

Rubber 3 1.16 

Shipyard 5 1.94 

Textile  13 5.04 

Trade 9 3.49 

Transport 2 0.78 

Grand Total 258 100.00 

 

The Table shows that the leading companies in Japan are mostly in the financial / banking sector 

(18.99%), the chemical industry (10.85%), the food industry (10.08%) and electronics sector (8.53%). 

 

3. Result 

3.1 Network and ERGM Results 

A 3a. and 3b. Figures show the network morphology of Kerietsu groups. Figure 3 shows the complete 

network in which we can see the Kerietsu groups’ positions as also the positions of the blocks and 

clusters, while network Figure 3b depicts only the network of companies belonging to Kerietsu groups. 

Both networks are by definition unidirectional. Table 4 summarizes the basic network statistics for the 

two networks. 

Figures 3 and 4 - Kerietsu groups and block model network and Kerietsu groups network 

 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

53 
Published by SCHOLINK INC. 

 

Figure 3. Kerietsu Groups and Block Model Network (N=283) 

 

 

Figures 4. Kerietsu Groups Network (N=117) 

 

Table 4. Network Statistics of the Two Networks 

 Kerietsu groups and block model 

network (entire network) 

Kerietsu groups network 

Node number 283 117 

Edge (tie) number 369 111 

Average edge degree 1.304 0.949 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

54 
Published by SCHOLINK INC. 

Network density 0.009 0.016 

Modularity 13 6 

Average path 1 1 

Number of components 1 6 

 

From the above statistics and morphology, it can be clearly seen that the six Kerietsu, which comprise a 

total of 117 leading Japanese companies, decompose into 6 well-separated components with high 

network density. Each Kerietsu encompasses companies from a very wide spectrum of economic 

sectors, from the banking system to heavy industry. 

However, it is better to move on to an analysis of the entire network, which shows that Kerietsu 

networks play a much smaller role in the network of all leading companies. Statistics and morphology 

indicate that these corporate organizations are embedded in an economic network characterized by a 

much larger number of edges and a much lower network density. They form a completely connected 

network with 13 well-separable subgroups. An important result is that the role of Kerietsu corporate 

forms in the economic network has been taken over by Groups, in which members have an even less 

close network relationship with each other. Thus, it is not unlikely that the Mitsui Group has the highest 

degree (33) in the entire network. However, we still find Kerietsu groups in the most central positions 

like Daiichi Kangyo keiretsu (25) Sanwa keiretsu (24) and Fuyo keiretsu (19). We can find the City 

Central Banks (19) and other groups such as the Mitsubishi Group (18) and the DKB and Sanwa Group 

(17) in the most centric positions of the network. 

Let us move on to the ERGM analysis itself, based on the Kerietsu groups and block model network. 

The network represents the entire network of connections between large Japanese corporations, in 

which organizations connect with each other for market dominance. ERGM analysis helps us to explore 

along which exogenous and endogenous variables this market network has evolved. There are many 

possible competing explanations for the evolution of the network structure, but if we examine only one 

process at a time, we may overestimate its role. With the help of ERGM, we can deduce the presence of 

an independent tendency for a particular configuration to appear, or the tendencies whose co-presence 

shapes the network. 

The parameter estimates of the ERGM model were compared with 500 Markov Chain Monte Carlo 

(MCMC) network simulations. MCMC simulations created networks with the same number of nodes 

and densities as the observed network. In comparison, the MPnet software generated parameter 

estimates of the model, which indicate the strengths and direction of the network samples (endogenous 

and exogenous). Table 5 shows the parameter estimates for the Japanese economic network. Significant 

parameters are marked with an asterisk (*). A positive (negative) estimate indicates a larger (lesser) 

configuration than expected in the network, taking into account the other effects of the model. The 

magnitude of the parameter estimates cannot be compared directly along the different effects, because 

the scaling of the statistics is different. 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

55 
Published by SCHOLINK INC. 

Table 5. ERGM Parameter Estimates for the Japanese Economic Network 

Network effects  Estimates - SE 

Edge effect (Arc)  -0.091 

2-Star (S2) 

 

74.607* 

3-Star (S3) 

 

218.904* 

4-Star (S3)  781.862* 

5-Star (S3)  3320.958* 

Triad effect 

 

-1.692  

4-Cycle 

 

88.190* 

Isolate  -5.393* 

Hub 

 

11.228* 

Multiple 2-path 

 

71.058* 

Transitivity (Transitive Multiple  

2-path)  

-1.701 

Alt-Edge-Triangle (AET) 

  

-1.643 

Attribute effects 

(exogenous - the blue circle indicates the node that can be characterized by the attribute) 

Homophily (econ. sector)  2.700 

* = the parameter estimate is twice the absolute value of the standard error, which means that the effect 

is significant. 

 

Based on the ERGM, the network can be characterized by a negative edge effect (-0.091); however, this 

effect is not significant. In other words, it is not the pair connections that play a decisive role in the 

structure of the network. What does play an important role in the formation of the network is whether 

or not the firm connects to the actors in the central positions, such as the Kerietsu groups and the block 

groups. We see this in the network-forming power of the 2, 3, 4, and even 5-star shapes, which in each 

case have positive and significant values. All this means that network centers play a crucial role in the 

formation of the network and are surrounded by a multitude of less inter-connected companies. The 

same result is supported by the hub effect, i.e., the positive and significant effect of the degree 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

56 
Published by SCHOLINK INC. 

distribution (11.228 *). 

The strong centralization effect of the network is also supported by the negative and non-significant 

value of the triad effect, which means that the triple closed connections do not substantially shape the 

network. This is somewhat contradicted by the positive and significant role of quartet closed 

connections (88,190*), which can be explained by the fact that quadruple connections are formed quite 

often between hubs (groups, Kerietsu, blocks and clusters), which also indicates the strong network 

centralization. 

The multiplied 2-way effect is also positive and significant, indicating that we find more 2-ways, due to 

the other effects of the model, in the measured network, than we might expect, i.e., connections with 

several central nodes forming the same companies are related with each other. However, the transitivity 

effect (transitive triad effect) is negative and not significant, suggesting that these 2-pathways do not 

close, but condense around specific nodes. We see the same lack of closure among the actors in the fact 

that we do not get a significant value in the case of alternative edge triangles either. 

After the endogenous effects, let us move on to the exogenous effects, which in this case cover the 

attribute effect, i.e., the extent to which the economic sector influences the development of the network. 

We can see that the value is positive (thus influencing it in theory), but not significant, i.e., it does not 

play a decisive role with the other endogenous effects. 

 

4. Discussion 

I have analyzed the Japanese network economy in my paper, because in more and more countries, the 

system of relations between the leading market companies and financial institutions is constantly 

intensifying. In other words, we can see a tendency of monopolization in the capitalist economy. I have 

illuminated the effects of this trend through an economic system that has operated in a similar 

centralized relationship system for many decades. Starting with a review of the historical development 

of this economy, I have interpreted the available data with an ERGM network analysis. 

From the ERGM analysis, we have seen that we can still analyze the network of large Japanese 

companies with highly endogenous effects resulting from network connections. This is not a 

coincidence, because, as has been raised several times in the historical part of my article, the Japanese 

economy relies heavily on resource transfers and on different corporate circles. All this means that the 

existence of a relationship of trust and dependency between corporate actors is decisive in the Japanese 

economy to this day, and this network effect is likely to persist for a long time due to the inertia of the 

system. 

It can also be seen in the Keiretsu and group formations that the dominant relationships extend from the 

core - from the most important banks, commercial companies or heavy industry companies - to the 

extreme periphery, and we encounter less and less edge density, moving from the core to the periphery. 

While this will of course help to improve the quality of information flows, positioning the common 

market goals and speed up production and logistics processes, at the same time, as we have seen in the 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

57 
Published by SCHOLINK INC. 

Japanese economy, it can also lead to a decline in innovation and new entrants or independent entrants, 

not to mention that such a structure is beneficial primarily for companies located in the core, and as we 

move towards the periphery, this ’positive’ impact diminishes. 

As we have seen since the 2010s, Japan has decided that this economic structure, which is based on 

centralized market relations, will not make the country competitive in the global market in the long run, 

and we can also see that this systemic change has proved true. However, the anomalies of a 

multi-decade system can be changed over only extremely long decades. For all these reasons, it would 

be important for many national economies to consider the fact that a centralized and inter-dependent 

economic system cannot be sustainable in the long run in the current global economy.  

 

References 

Agarwal, R., & Elston, J. A. (2001). Bank-firm relationships, financing and firm performance in 

Germany. Economics Letters, 72(2), 225-232. https://doi.org/10.1016/S0165-1765(01)00427-X 

Brewer, E., Genay, H., & Kaufman, G. G. (2003). Banking relationships during financial distress: the 

evidence from Japan. ECONOMIC PERSPECTIVES-FEDERAL RESERVE BANK OF CHICAGO, 

27(3), 2-18. 

Chan-Lau, J. A. (2002). Corporate restructuring in Japan: An event-study analysis. Japan and the World 

Economy, 14(4), 367-377. https://doi.org/10.1016/S0922-1425(02)00031-2 

De Masi, G., Fujiwara, Y., Gallegati, M., Greenwald, B., & Stiglitz, J. E. (2011). An analysis of the 

Japanese credit network. Evolutionary and Institutional Economics Review, 7(2), 209-232. 

https://doi.org/10.14441/eier.7.209 

Driffield, N., & Munday, M. (2000). Industrial performance, agglomeration, and foreign manufacturing 

investment in the UK. Journal of International Business Studies, 31(1), 21-37. 

https://doi.org/10.1057/palgrave.jibs.8490897 

Driffield, N., & Taylor, K. (2000). FDI and the labour market: A review of the evidence and policy 

implications. Oxford Review of Economic Policy, 16(3), 90-103. 

https://doi.org/10.1093/oxrep/16.3.90 

Driffield, N. (2001). The impact on domestic productivity of inward investment in the UK. The 

Manchester School, 69(1), 103-119. https://doi.org/10.1111/1467-9957.00237 

Driffield, N., & Girma, S. (2003). Regional foreign direct investment and wage spillovers: Plant level 

evidence from the UK electronics industry. Oxford Bulletin of Economics and Statistics, 65(4), 

453-474. https://doi.org/10.1111/1468-0084.t01-1-00057 

Gerlach, M. L. (1992). The Japanese corporate network: A block model analysis. Administrative 

Science Quarterly, 105-139. https://doi.org/10.2307/2393535 

Grabowiecki, J. (2006). Keiretsu groups: Their role in the Japanese economy and a reference point (or 

a paradigm) for other countries. Institute of Developing Economies, Japan External Trade 

Organization. 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

58 
Published by SCHOLINK INC. 

Guillaume, J. L., & Latapy, M. (2004). Bipartite structure of all complex networks. Information 

processing letters, 90(5), 215-221. https://doi.org/10.1016/j.ipl.2004.03.007 

Hadley, E. M. (2015). Antitrust in Japan. Princeton University Press. 

https://doi.org/10.1515/9781400872053 

Imai, K. I. (1987). The corporate network in Japan. Japanese Economic Studies, 16(2), 3-37. 

https://doi.org/10.2753/JES1097-203X16023 

Kanda, H. (1998). Comparative corporate governance country report: Japan. In K. J. Hopt, H. Kanda, 

M. J. Roe, E. Wymeersch, & S. Prigge (Eds.), Comparative Corporate Governance: The State of 

the Art and Emerging Research. Oxford, Clarendon Press,  

Kimino, S., Saal, D. S., & Driffield, N. (2007). Macro determinants of FDI inflows to Japan: An 

analysis of source country characteristics. World Economy, 30(3), 446-469. 

https://doi.org/10.1111/j.1467-9701.2007.01001.x 

Lincoln, J. R., Gerlach, M. L., & Takahashi, P. (1992). Keiretsu networks in the Japanese economy: A 

dyad analysis of interoperate ties. American Sociological Review, 561-585. 

https://doi.org/10.2307/2095913 

Lincoln, J. R., & Gerlach, M. L. (2004). Japan’s network economy: Structure, persistence, and change. 

Cambridge University Press. https://doi.org/10.1017/CBO9780511584442 

Lusher, D., Koskinen, J., & Robins, G. (Eds.). (2013). Exponential random graph models for social 

networks: Theory, methods, and applications (Vol. 35). Cambridge University Press. 

https://doi.org/10.1017/CBO9780511894701 

Marotta, L., Micciche, S., Fujiwara, Y., Iyetomi, H., Aoyama, H., Gallegati, M., & Mantegna, R. N. 

(2015). Bank-firm credit network in Japan: an analysis of a bipartite network. PloS one, 10(5), 

e0123079. https://doi.org/10.1371/journal.pone.0123079 

Morck, R., & Nakamura, M. (1999). Banks and corporate control in Japan. The Journal of Finance, 

54(1), 319-339. https://doi.org/10.1111/0022-1082.00106 

Ogawa, K., & Wan, J. (2007). Household debt and consumption: A quantitative analysis based on 

household micro data for Japan. Journal of Housing Economics, 16(2), 127-142. 

https://doi.org/10.1016/j.jhe.2007.02.001 

Peltomäki, M., & Alava, M. (2006). Correlations in bipartite collaboration networks. Journal of 

Statistical Mechanics: Theory and Experiment, 2006(01), P01010. 

https://doi.org/10.1088/1742-5468/2006/01/P01010 

Polányi, K. (1944; 2001). The Great Transformation: The Political and Economic Origins of Our Time. 

Boston. Beacon Press 

Sneppen, K., Maslov, S., & Zaliznyak, A. (2004). Detection of topological patterns in complex 

networks: correlation profile of the internet. Physica A: Statistical Mechanics and its Applications, 

333, 529-540. https://doi.org/10.1016/j.physa.2003.06.002 

Spiegel, M. M., & Yamori, N. (2003). The impact of Japan's financial stabilization laws on bank equity 



www.scholink.org/ojs/index.php/ape              Advances in Politics and Economics                 Vol. 4, No. 2, 2021 

59 
Published by SCHOLINK INC. 

values. Journal of the Japanese and International Economies, 17(3), 263-282. 

https://doi.org/10.1016/S0889-1583(03)00006-6 

Sterken, E., & Tokutsu, I. (2002). What are the determinants of the number of bank relations of 

Japanese firms?. University of Groningen. 

Takeda, Y., Kajikawa, Y., Sakata, I., & Matsushima, K. (2008). An analysis of geographical 

agglomeration and modularized industrial networks in a regional cluster: A case study at Yamagata 

prefecture in Japan. Technovation, 28(8), 531-539. 

https://doi.org/10.1016/j.technovation.2007.12.006 

Yonekura, S. (1985). THE EMERGENCE OF THE PROTOTYPE OF ENTERPRISE GROUP 

CAPITALISM—THE CASE OF MITSUI—. Hitotsubashi journal of commerce and management, 

20 (20), 63-104. 

Yoshino, N., & Taghizadeh-Hesary, F. (2014). An analysis of challenges faced by Japan’s economy and 

Abenomics. The Japanese Political Economy, 40(3-4), 37-62. 

https://doi.org/10.1080/2329194X.2014.998591 

Wang, P., Robinson, G., Pattison, P., & Koskinen, J. (2009) MPNet: Program for the simulation and 

estimation of exponential random graph models. Melbourne School of Psychological Sciences, 

The University of Melbourne. 

Watts, D. J. (1999). Networks, dynamics, and the small-world phenomenon. American Journal of 

Sociology, 105(2), 493-527. https://doi.org/10.1086/210318  

World Bank. (2020). Retrieved from https://databank.worldbank.org/home.aspx 

 

Notes 

Note 1. CONCOR was a disaggregative clustering algorithm: it iteratively divided the population into 

several smaller sets. 

Note 2. I use the terminology of capitalism as a collective term in my article. Of course, the majority of 

nation-states in the twentieth century can be placed in an extremely wide spectrum in this aspect, 

encompassing a wealth of forms from state-capitalist economies, through the monopoly-capitalist 

economy to the classical capitalist economies. 

Note 3. The Big Six terminology is used to denote the leading groups of companies in the Japanese 

economy,namely Sumitomo, Mitshubishi, Fuji, Sanwa, Mitsui and Daiichi Kangyo Group of 

Companies. 

 

 

 


