As established in much of the foreign direct investment (FDI) literature, multinational enterprises (MNEs) often borrow to pay for a portion of a newly established foreign subsidiary (e.g. Buch 2014, Yan 2018, Bilir 2019). Two channels exist through which financial frictions may affect the firm's investment ability and ownership structure decision. First, investors face a demand-side constraint, or ‘borrowing channel’, that constrains its borrowing ability because firms typically collateralise the amount borrowed. Difficulties in doing so, or other balance sheet-related issues (e.g. high debt), raise the cost of borrowing, limiting the firm’s ability to finance its investment projects.
Second, a ‘lending channel’ arises from weak credit markets, typically resulting from decreased institutional health through which frictions facing banks and other lending institutions impact their ability to provide loans to MNE borrowers.
Previous theoretical and empirical research shows that financial frictions in both channels negatively impact investment totals and affect ownership choice decisions (whole ownership or joint venture). For many investors, a joint venture arrangement may be the only possible method of establishing a foreign subsidiary. This result is in contrast to earlier ownership choice studies that assumed MNEs could establish wholly owned subsidiaries and that the ownership choice decision was based on factors such as solving ex-post incentive problems and contractual issues.
A drawback in this literature is that data tying individual loans and collateral requirements to specific FDI projects often do not exist. To ease identification issues, assumptions are made regarding who serves as the MNE’s main lending institution (‘main bank’) and the firm limits its borrowings to this institution. However, the main bank as the sole lender is generally untrue: the Nikkei NEEDS Corporate Borrowings from Financial Institutions Database of matched bank-firm loan data indicates that approximately 70% of Japanese firms borrowed from more than one bank between 1985–2019.
In a recent paper (Ito et al. 2023), we take a different approach to modelling the lending channel. Rather than assuming loan-investment links or a lender’s financial health, we explore lending constraints through a key characteristic of Japanese firm-level data that identifies the degree to which these lenders serve as shareholders of their MNE clients. In identifying where these relationships exist and the extent to which the bank lender regulates its client's investment behaviour, we explore a different line of heterogeneity among investing firms while addressing a different lending-channel constraint faced by investing firms.
We chose Japanese MNE FDI behaviour due to Japan’s prominent role as a major FDI source and the availability of Japanese firm-level FDI and balance sheet data that ties individual MNE financial health to their investment activities. In contrast to most other major FDI source countries, Japanese firms often have banks or other lending institutions as their major owners. Additionally, and unique to Japan, banks often have personnel sitting on the MNE’s board of directors and (to varying degrees) participating in its operational decisions. As such, the main bank as a top shareholder can not only exert power over a firm but also serve as a conduit of investment information (e.g. Inui et al. 2015, Degryse et al. 2009). As the shareholding bank’s profit is directly related to the firm’s profit, the main bank has the incentive to provide financial and informational support for the firm’s profitable FDI projects.
Aoki and Patrick (1994) argue that close Japanese bank-firm ties have helped solve agency problems and asymmetric information issues. Hoshi et al. (1991) find that for investment, firms with close main bank ties are much less sensitive to liquidity constraints than firms raising their capital through more arms-length transactions. While this implies that the main banking relationship can minimise agency and information problems, these informational asymmetries still exist and can be larger for outward FDI. Therefore, the main bank’s risk aversion can serve to lower a MNE’s ownership ratio of a foreign subsidiary. Similarly, MNE subsidiaries established in countries where the parent’s main bank shareholder already operates banking subsidiaries should see information asymmetries alleviated, resulting in higher ownership percentages for its newly established affiliates. This mimics the ‘follow the customer’ FDI literature, where manufacturing MNEs often choose foreign hosts in which their main bank has already established subsidiaries (e.g von der Ruhr and Ryan 2005). In this case, FDI information issues regarding the host country are less prominent, as the bank’s foreign branch has information on the local market that can be used to reduce the cost/uncertainty of MNE investment there. With this information flow, investment there would be considered a better lending risk, and the MNE more likely can establish a wholly owned subsidiary there.
We examine FDI by Japanese publicly held MNEs into manufacturing affiliates in 58 hosts – 29 OECD and 29 non-OECD countries – between 1989 and 2016. For affiliate-specific data, we employ Toyo Keizai Inc.’s firm-level Overseas Japanese Companies Data (OJC), which covers nearly the entire universe of Japanese foreign affiliates. The OJC lists each affiliate firm’s name, geographic location, establishment year, its parent firms (whether they be Japanese, local, or third country), and parent ownership percentages. Detailed non-consolidated corporate financial records for each MNE are found in the Development Bank of Japan’s Corporate Financial Data Bank (DBJ data).
Our key variable of interest – the MNE’s ownership ratio of its foreign subsidiaries – is bounded between 0 and 1, so we employ a fractional logit model regression model. For each MNE parent, we a priori assume the firm’s Total Factor Productivity and its Intangibles Ratio – total intangible fixed assets as a percentage of total fixed assets – to be positively related to ownership share, while Debt Ratio and Top Bank Ratio – the largest shareholder bank’s ownership ratio – are assumed to be negatively related to affiliate ownership share. We include year, host-country, and parent-industry fixed effects, and all explanatory variables are calculated with a two-year lag to account for the time between the investment decision and the affiliate’s operational start.
Figure 1 provides the average marginal effect on foreign affiliate ownership in three scenarios: using all 58 host countries, into just OECD hosts, and into non-OECD hosts. As expected, TFP typically positively and statistically significantly affects affiliate ownership share, likely driven by the significant coefficient for the non-OECD hosts. Essentially, firms must be more productive when investing in non-OECD hosts. Debt Ratio is unsurprisingly negative and statistically significant. Importantly, we find that Top Bank Ratio serves to significantly reduce the MNE’s affiliate ownership share in both the whole sample and non-OECD hosts. Investing in non-OECD hosts is often viewed as riskier than in OECD hosts, and the MNE’s bank parent’s risk aversion serves to lower investment risk by lowering the MNEs’ affiliate ownership shares.
Figure 1
Research shows previous FDI experience can prove valuable in establishing an optimal future affiliate ownership structure. Experienced parent firms can easily obtain external financing by providing existing foreign subsidiaries with collateral. In such cases, the experienced parent firm’s financial constraints and productivity can become less important determinants of its foreign subsidiaries’ ownership structures. Considering this possibility, we create a subsample covering the MNEs’ initial FDI projects in each host country. Figure 2 examines the ownership share of all initial FDI projects, as well as the likelihood of establishing either a wholly owned or majority-owned affiliate. Debt Ratio negatively impacts all investment choices, and it is not unsurprising that for firms that can establish wholly owned subsidiaries, the Top Bank Ratio does not affect investment likelihood. In contrast, however, Top Bank Ratio does suppress ownership share and the likelihood of majority ownership.
Figure 2
For many Japanese companies, banks hold a substantial ownership percentage and are often the largest shareholders. Firms with higher bank ownership are often viewed as less susceptible to more severe financial constraints. Nevertheless, information asymmetries between firms and banks are greater for foreign than domestic investments. How banks influence firms’ ownership decisions related to foreign subsidiaries cannot be determined without empirical analysis. We create an indicator variable, Bank as Top Owner, which takes the value of 1 if the largest shareholder of the parent firm is a bank. Figure 3 shows that parent firms with banks as the largest owners tend to have lower ownership ratio of their foreign subsidiaries and tend to choose minority- rather than majority-owned foreign subsidiaries. The results suggest that banks, as shareholders, prefer risk-averse FDI with a lower ownership ratio of foreign subsidiaries.
Figure 3
Figure 4 examines the impact of the MNE’s shareholding banks’ FDI behaviour. In host countries where the MNE’s main banks have foreign subsidiaries, MNEs tend to increase their ownership and lower non-Japanese ownership of their foreign subsidiaries. The size of the main bank’s overseas network is also associated with higher MNE affiliate ownership and lower non-Japanese partner affiliate ownership share. This supports the ‘follow the customer hypothesis’ that MNEs are more likely to follow their lender abroad, as their presence tends to ease host country informational asymmetries and investment costs.
Figure 4
Conclusion
The COVID-19 pandemic reminded us of global supply chain fragility. Firms seeking to integrate their supply chains and minimise current bottleneck issues with contract suppliers must create or acquire subsidiaries in new locations and/or industry sectors outside of their traditional competencies. Doing so leads to greater MNE investment risk-taking. Given the risk-averse nature of the firm’s main banks and shareholders, these investments likely face significant funding scrutiny, creating sizeable roadblocks to supply chain reorganisations. Recognising the importance of these supply chains to national welfare, our results indicate that national governments such as Japan, which are wishing to assist in MNE foreign affiliate network reorganisation, may consider simultaneously working to improve shareholding banks’ health and foreign investment activities.
Editor’s note: The main research on which this column is based (Ito et al. 2023) first appeared as a Discussion Paper of the Research Institute of Economy, Trade and Industry (RIETI) of Japan.
References
Aoki, M and H Patrick (1994), The Japanese Main Bank System: Its Relevance for Developing and Transforming Economies, Oxford University Press.
Bilir, L K, D Chor, and K Manova (2019), “Host-country Financial Development and Multinational Activity,” European Economic Review 115: 192–220.
Buch, C M, I Kesternich, A Lipponer, and M Schnitzer (2014), “Financial constraints and foreign direct investment: firm-level evidence,” Review of World Economics 150(2): 393–420.
Chaney, T, D Sraer, and D Thesmar (2012), “The Collateral Channel: How Real Estate Shocks Affect Corporate Investment,” American Economic Review 102(6): 2381–2409.
Degryse, H, M Kim, and S Ongena (2009), Microeconometrics of Banking: Methods, Applications, and Results, Oxford University Press.
Gan, J (2007a), “Collateral, debt capacity, and corporate investment: Evidence from a natural experiment,” Journal of Financial Economics 85(3): 709–734.
Gan, J (2007b), “The real effects of asset market bubbles: Loan-and firm-level evidence of a lending channel,” The Review of Financial Studies 20(6): 1941–1973.
Gibson, M S (1995), “Can bank health affect investment? Evidence from Japan,” Journal of Business 68(3): 281–308.
Hoshi, T, A Kashyap, and D Scharfstein (1991), “Corporate structure, liquidity, and investment: Evidence from Japanese industrial groups,” The Quarterly Journal of Economics 106(1): 33–60.
Inui, T, K Ito, and D Miyakawa (2015), “Overseas Market Information and Firms’ Export Decisions,” Economic Inquiry 53: 1671–1688.
Ito, T, M Ryan, and A Tanaka (2023), “Partial ownership, financial constraint, and FDI,” RIETI Discussion Paper Series 23-E020.
Klein, M W, J Peek, and E S Rosengren (2002), “Troubled Banks, Impaired Foreign Direct Investment: The Role of Relative Access to Credit,” American Economic Review 92(3): 664–682.
Raff, H, M Ryan, and F Stähler (2018), “Financial frictions and foreign direct investment: Evidence from Japanese microdata,” Journal of International Economics 112: 109–122.
von der Ruhr, M and M Ryan (2005), ““Following” or “attracting” the customer? Japanese Banking FDI in Europe,” Atlantic Economic Journal 33(4): 405–422.
Yan, B, Y Zhang, Y Shen, and J Han (2018), “Productivity, Financial Constraints and Outward Foreign Direct Investment: Firm-level Evidence,” China Economic Review 47: 47–64.
