The old saying ‘no man is an island’ applies equally to firms. The productivity of individual firms depends on their interactions with their suppliers and customers. It is important for a firm to search for appropriate transaction partners and construct relationships.
Moreover, the forms of these linkages matter. The linkages tend to occur within geographically limited areas due to various search costs. To expand networks, policymakers have tried to promote firms to construct relationships with many other firms by reducing the cost. This has taken the form of providing a ‘matching’ service. The same end is pursued by private firms, namely, wholesalers who seek to expand transaction networks as intermediaries.
Theoretical view of wholesalers’ roles
The role of wholesalers as intermediaries has been studied in previous studies of international trade. Ahn et al. (2011) and Bernard et al. (2011) compare firms that export directly and those that do so indirectly through wholesalers. They find that less productive firms tend to rely on wholesalers for exporting, suggesting the existence of costs directly involved with transacting that cannot be overcome by unproductive or small firms. Most related existing literature, however, only considers one side of the transaction (i.e. the seller side) and focuses on how firms use wholesalers for distribution (i.e. exporting). Furthermore, the difference in wholesalers’ roles between distribution and procurement has not yet been explored in detail.
When a manufacturing firm sells directly to other manufacturing firms, it searches for buyers and takes the risk of not being paid. When a manufacturing firm procures an input directly from other manufacturing firms, it searches for sellers and takes the risk regarding the quality of the input. Therefore, we expect that the role of wholesalers for distribution is different from that for procurement. We investigate wholesalers’ roles in transaction networks, considering both sides of transactions.
Data and methodology
We use unique data on Japanese domestic transactions compiled by Tokyo Shoko Research, Ltd. The data include 800,000 firms and 4 million firm-level transaction relationships. Similar to international trade, it is found that only a few hub firms transact with remote firms in this domestic network, suggesting the existence of geographical friction of transactions (Saito 2013). It is possible that wholesalers mitigate geographical friction of transactions and contribute to geographical expansion of domestic transaction networks.
Starting from an overview on a relative position of wholesalers to that of manufacturing firms in the transaction networks, we calculate standard centrality indices used in network analysis, i.e. closeness and ‘betweenness’ indices. We find that the betweenness index for wholesalers is larger than that for manufacturing firms, although the closeness indices are similar. The large value of betweenness indicates an essential role in connecting other firms by wholesalers.
We then compare manufacturing firms transacting directly with those connected indirectly through wholesalers, focusing on transaction networks for intermediate manufacturing goods. We use a probit model to examine which characteristics are associated with a firm’s likelihood to use wholesalers. We perform such analyses separately for manufacturers’ use of wholesalers for distribution and that for procurement.1 Furthermore, we examine the distance between transaction partners to see whether wholesalers are located near to their manufacturing buyers or sellers. By locating themselves close to manufacturer buyers, it is possible that wholesalers collect information on the manufacturers’ needs as well as their financial states to reduce transaction costs and limit risks.
Main results and implications
We find a clear asymmetry in the characteristics of buyer- and seller-side firms that tend to rely on wholesalers.
- The likelihood that a firm uses wholesalers increases with smaller buyer-side firms and larger seller-side firms.
- In addition, the likelihood increases if buyer-side firms are located in less dense areas and seller-side firms are located in more dense areas.
- In terms of distance between transaction partners, we find that wholesalers tend to be located closer to their manufacturing buyers and further from their manufacturing sellers than manufacturers are to their direct manufacturing partners.
Such asymmetry has not been found in the existing literature.
A possible interpretation of our findings is as follows. Wholesalers play different functions depending on the size and geographical location of their manufacturing partners. It is possible that small firms in rural areas tend not to have enough credit reputation and negotiating power. Therefore, they rely on wholesalers that buy directly from large firms in urban areas. By locating themselves near to manufacturing buyers, wholesalers can monitor their financial condition. Wholesalers mitigate geographical friction in transactions and accelerate indirect transactions between firms in rural and urban areas. On the other hand, large firms in urban areas procure directly from sellers, monitoring product quality by themselves, and they use wholesalers to distribute goods. Thus, wholesalers perform the role of distributors for large firms.
Ahn, J B, A K Khandelwal, and S J Wei (2011), “The Role of Intermediaries in Facilitating trade,” Journal of International Economics 84, 73-85.
Bernard, A, M Grazzi, and C Tomasi (2011), “Intermediaries in International Trade: Direct versus Indirect Modes of Export,” NBER Working Paper No. 17711
Okubo, T, Y Ono, and Y U Saito (2014), "Roles of Wholesalers in Transaction Networks," RIETI Discussion Papers series, 14-E-059, October.
Saito, Y U (2013), "Role of Hub Firms in Geographical Transaction Network," RIETI Discussion Papers series, 13-E-080, September.
 As characteristics of transaction partner, we use those of indirect partners when the firm uses wholesalers and those of direct partners otherwise. When several firms are connected by the wholesaler, we calculate the average values of their characteristics.