1
The effect of firm internationalization, market development, and Financial
crisis on capital structure. Evidence from Western European and Eastern
European firms
University of Groningen Faculty of economics and business
Masters Finance
Student: Emmanuel Vuyof Atsimbom
S3850838
Supervisor: Dr. Swarnodeep Homroy
2 Abstract:
3 JEL classification: G32, G34, F32, F36, F63 and H63.
Keywords: Corporate Governance, Capital structure, Leverage ratios, internationalization, financial development, Developing countries and developed countries, Eastern Europe, Western Europe, European Union, Financial crisis.
Student number: S3850838
4
Contents
1. Introduction ... 5
2. Empirical review ... 10
2.1. Firm Internationalization and capital structure ... 10
2.2. Institutional development, Financial development, and capital structure ... 12
2.3. The global financial Crisis ... 16
3. Data and methodology ... 19
3.1. Methodology ... 19
3.2. Data ... 27
3.3. Descriptive statistics ... 28
4. Regression Results ... 34
4.1. Regional development, Firm internationalization, and capital structure ... 34
4.2. Eastern European firms: Multinational and Domestic firms ... 38
4.3. Financial Crisis, Internationalization, and capital structure ... 40
4.4. Robustness Checks ... 43
5. Conclusion ... 44
6. Limitations ... 47
Bibliography ... 48
5
1. Introduction
Regional legal, political, and financial institutions are key determinants of how the average firm adjusts its capital structure (leverage), and variations in the quality of these institutions across Eastern Europe and western Europe influence differences in the capital structure choices that firms between these geographic regions make. Countries with weaker institutional proxies issue and use less debt and equity ( (Pietro, Palacin-Sanchez, & Roldan, 2018; Oztekin, 2015). Firms in countries with poor financial-institutional and poor legal and political-institutional compliance demonstrate higher levels of information asymmetry. This leads to an increase in the cost of equity and the cost of debt, resulting in the use of less debt and equity in such regions (Huang & Shen, 2015). The integration of developing and developed countries provides financial opportunities to firms in developed countries through the globalization (internationalization) of trade and transactions, and the flow of investment and capital. Through internationalization, developing country firms are exposed to higher financial markets with lower costs of financing, thus exposing developing country firms to more debt and equity financing (Gonenc & Haan, 2014). Internationalization exposes developing countries’ firms to the same quality of financing that firms in developed countries are exposed to (Agarwal, Agarwal, & Agarwal, 2006; Lothian, 2006). The accession of Eastern Europe into the EU follows a recent trend of integration between developing and developed countries, where Eastern European countries are the developing countries and Western European countries are the developed countries. According to Kwok & Reeb (2000), when countries based on the less stable economies internationalize to countries based in more stable economies, they are exposed to lesser risk. As a result, they become exposed to more external financing and thus engage in higher leverage compared to their domestic counterparts. This results in capital flow from developed countries into developing countries (Gonenc & Haan, 2014; Vliegenthart & Horn, 2007).
6 decrease in the leverage of the companies (Kwok & Reeb, 2000). On the other hand, Canadian multinational firms with exposure to international bond markets have been found to use higher long-term debt than their domestic counterparts. The global expansion creates diversification of profitability opportunities for these firms through their presence in different country markets. The diversification of risks occurs because firms’ presence in multiple countries reduces operational risks for multinational companies (Mittoo & Zhang, 2008). Domestic companies do not have the opportunity to exploit these advantages. As a result of less operational risks and fewer agency costs of capital, multinational companies can assume more financial risk, and therefore increase their leverage ratios (Mittoo & Zhang, 2008; Lindner, Klein, & Schmidt, 2018). Based on the above-mentioned background, this study aims to answer the following two questions:
1) What is the effect of firm internationalization and level of country development on the capital structure of Western European firms and Eastern European firms (EU firms)?
2) Is there a difference in the capital structure of Eastern European multinational firms when compared to the capital structure of Eastern European domestic firms after the accession of Eastern Europe into the European Union (EU)?
7 Weibenner, 2009; Demirguc_Kunt, Peria, & Tressel, 2020). This background leads to the following third research Question:
3) What is the effect of the credit crisis on the capital structure of Eastern European firms and Western European firms?
The EU is a suitable and unexplored area for this analysis because, on the one hand, all the countries in the EU share the same economic market environment. On the other hand, the European Union shows regional disparities in the level of country development between Eastern European countries and Western European countries. This results in differences in institutional factors between the countries, opening the possibility for a detailed analysis of the key differences in characteristics between both regions (Demirguc-Kunt & Maksimovic, 1999; Vliegenthart & Horn, 2007; Gonenc & Haan, 2014). Examining these relationships using Eastern European firms provide evidence from a developing market perspective. Eastern Europe (developing countries) is characterized by less developed and sophisticated markets and inefficient capital markets and engages more in debt financing than equity financing. Developing countries depend on an institutionally bank-based financial system as their primary source of financing, while developed countries depend on the effective allocation of capital through market forces (Classens & Yurtoglu, 2013). Also, Eastern Europe has a high level of information asymmetry, high debt to equity ratios, underdeveloped legal markets, and economic and political instability in this region. The financial markets in this region are also less liquid. The level of liquidity in financial markets causes Eastern European financial markets to be very volatile. This high level of volatility further hinders the development of the markets in this region. Furthermore, Eastern European markets have less access to sources of financing, limited financial market development, and low level of institutional ownership. These unique characteristics of the Eastern European developing market could influence the corporate capital structure choices that firms in this region make, which could differentiate the financial structure of firms in Eastern Europe from the financial structure of firms in Western Europe (ElBannan, 2017; Vliegenthart & Horn, 2007; Tonoyan, Strohmeyer, Habib, & Perlitz, 2010).
8 Filatotchev, & Marshall, 2010; Akhtar, 2005). This paper contributes to filling this gap by following recent calls to further examine this area of corporate governance study to provide more analytical evidence (Zao, Luo, & Suh, 2004; Kirca & Yaprak, 2010). The first contribution of this paper to the existing literature is to provide evidence from a prominent economic region in the world that is characterized by differences in regional economic development, but at the same time share the same economic market environment. Eastern European capital market and financial institutional systems have different characteristics compared to those in western Europe (Gonenc & Haan, 2014; Vliegenthart & Horn, 2007). Second, the regional institutional differences in the EU provides a unique setting to test the impact of the financial crisis on the capital structure choices of developed countries and compare this to the capital structure of developing countries in the same economic market environment. However, this comparison becomes challenging because the availability of data is poor and much more limited for developing countries when compared to developed countries. Developed country firms provide diverse observations and variables suitable for this analysis. Third, no extant paper analyses the effect of both firm internationalization and the financial crisis, and their effect on firms’ capital structure choices for Eastern and Western European firms. Thus, providing additional evidence on how developing and developed country firms in the same economic market environment change their choice of external financing during the financial crisis is an important contribution in this area of study. Given that the EU was severely impacted by the financial crisis in the form of the European sovereign debt crisis, data collected from the European economic market environment is appropriate for this study.
Overall, this paper shows contradicting results on the effect of firm internationalization on firms’ leverage. On the one hand, firm internationalization does not increase leverage for Eastern European firms and Western European (EU)firms. On the other hand, a subsample analysis of Eastern European firms shows that internationalization has significant positive effects on the leverage choices of Eastern European firms. this confirms the worries of inconclusive results in this area of study (Avarma, Hazak, & Mannasoo, 2011; Gonenc & Haan, 2014; Bist, 2018; Mittoo & Zhang, 2008). This paper also finds that firms in developed countries have a better capital structure than firms in developing countries. Furthermore, this paper shows that Multinational firms in developing countries significantly use more total debt, long-term and debt to equity, but less short-term debt than their domestic counterpart. Finally, this paper’s findings show that after the financial crisis, Eastern European firms and Western European firms in general use less total debt and long-term debt. The use of long-term debt increased after 2013 for firms in developed countries. This paper highlights the importance of considering regional institutional characteristics to ascertain the effect of internationalization in the firms' leverage.
10
2. Empirical review
Firm-specific characteristics and country-specific characteristics around which companies operate cause differences in the conditions through which firms can access external financing. These differences in access to external financing explain the various patterns of capital structure observed in firms. The tradeoff theory and the pecking order theory are two of the most important capital structure theories (Mukherjee & Mahakud, 2012) that explains the relationship between firm-specific characteristics (internationalization, size, profitability, tangibility, growth, etc.), firm institutional systems (geographical regions, country financial and economic development, etc.), and firms’ leverage (Pietro, Palacin-Sanchez, & Roldan, 2018; Gonenc & Haan, 2014). The pecking order theory is explained under the assumption of information asymmetry in the market that exists between managers of firms and (potential) outside investors. The theory predicts that firms use a hierarchical pattern to finance business operations beginning with internal financing and then external financing (debt and equity financing) (Myers, 1984; Myers & Majluf, 1984; Mukherjee & Mahakud, 2012). The tradeoff theory on the other hand suggests that a firms’ capital structure is optimal (Mukherjee & Mahakud, 2012).
2.1. Firm Internationalization and capital structure
11 corporate international activity and firm credit rating, their research finds that internationalization improves the credit rating of multinational firms and results in a lower cost of debt for multinational firms. Khurana, Martin, and Pereira (2006) find similar results in their study of the impact of financial development on firms' demand for liquidity. They find that financial constrain causes firms to avoid current external investments to use cash for future profitability and growth investments. Firms with no constraint to external funding can easily obtain external funding to invest in current profitable investments. They further argue that financial development makes it easier for international firms to access lower cost of external financing (Khurana, Martin, & Pereira, 2006; Gonenc & Haan, 2014).
Western Europe, through the EU and the accession process, has integrated Eastern Europe’s financial markets to its institutional and corporate governance image. This transition process allows free trade and free movement of financial institutions within a unified EU. Eastern European firms become exposed to the more dynamic and developed Western European market, exposing the Eastern European firms to the debt and equity capital of the developed Western European countries. Also, the Eastern European firms became exposed to the influx of Western European banking systems, which improved the financial environment and created more exposure to debt capital for the Eastern European firms. The effect of information asymmetry and agency costs is eliminated in these new (developing) Eastern European countries’ financial markets (Gonenc & Haan, 2014). Developed countries’ financial markets reduce the risk of firms not meeting their capital structure. Therefore, firms operating in countries with more developed financial markets are better able to access external funds (Gonenc & Haan, 2014; Mansi & Reeb, 2002)
12 the leverage choice between local and international companies that operate in the Baltic States, based on a sample size of 87000 company year observations. They find that multinational companies in the Baltic States are more exposed to attracting external financing as well as using internal group financing than their domestic counterparts. These discussions imply that firm internationalization improves firms’ use of external capital and that multinational firms in developing countries have better leverage compared with their domestic. when countries based on the less stable economies internationalize to countries based in more stable economies, they are exposed to lesser risk. As a result, they become exposed to more external financing and thus engage in higher leverage compared to their domestic counterparts (Kwok & Reeb, 2000). These Arguments suggest the following first and second hypothesis:
First hypothesis;
𝐻10: Firm internationalization does not result in a significant change in the capital structure of firms
𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐼𝑁𝑇= 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑁𝑂𝐼𝑁𝑇
𝐻11: Firm internationalization results in a significant increase in the capital structure of firms 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐼𝑁𝑇 > 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑁𝑂𝐼𝑁𝑇
Second hypothesis;
𝐻20: Due to firm internationalization, Eastern European multinational firms use similar leverage in their capital structure compared with Eastern European domestic firms.
𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐸𝑀𝐶 = 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐸𝐷𝐶
𝐻21: Due to firm internationalization, Eastern European multinational firms significantly use more leverage in their capital structure compared with Eastern European domestic firms.
𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐸𝑀𝐶 > 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐸𝐷𝐶
Where 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐼𝑁𝑇 proxy the capital structure resulting from firm internationalization, 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑁𝑂𝐼𝑁𝑇 proxy the capital structure of firms without internationalization, 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐸𝑀𝐶 proxy the capital structure of Eastern European multinational firms, and 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐸𝐷𝐶 proxy the capital structure of Eastern European Domestic firms.
2.2. Institutional development, Financial development, and capital structure
2.2.1 Background of Eastern Europe13 institutional underpinnings during the initial years of the transition process. These originated from legal, political, and institutional problems from the discredited Eastern European socialist models in the past. The leverage of firms depends on the institutional practices that determine the relationship between the firm and the stakeholders who finance the firm. The institutional transition process involves foreign direct investments, financial conditionality, institution building, and norm-setting. The EU was most influential in this transition which started after the Eastern European accession treaty of 1998. According to this 1998 treaty, new Eastern European member states must adapt to the EU institutional structure. The accession involved changing the logic of legal, political, and financial behaviors at the national level. By becoming part of the political structure and public policy, the EU influenced Eastern Europe’s financial environment to open and mimic Western Europe’s financial environment (Vliegenthart & Horn, 2007). Therefore through the accession treaty and before the Eastern European countries joined the EU in 2004, the EU has shaped the entire range of legal, political, and financial institutional policies of Eastern Europe (Vliegenthart & Horn, 2007). During the accession period, the EU influence on the institutional characteristics was broader than and deeper in scope than after Eastern Europe officially joined the EU. The EU stance to the accession requirements was particularly assertive and unyielding with regards to the accession and acquis conditionality of meeting the Copenhagen criteria to join the EU (Vliegenthart & Horn, 2007; Grabbe, 2003; Grabbe, 2001; Europa, 2020). Eastern European firms had been adjusting their capital structure to meet the changing institutional environment that was taking place during the accession period because of the push by the EU for the applicant countries to converge to the particular institutional models put in place, requiring applicants to stable and institutional market economies, and adherence to the political institution (Grabbe, 2001; Vliegenthart & Horn, 2007). Before 2004, the main economic priorities of the EU were for the liberalization and imposition of financial disciplines that led to the privatization of several financial institutions in Eastern Europe and increased the liquidity of financial assets in the region. This exposed firms in the region to opportunities to make better capital choices. Therefore the legal, political and financial institutional changes that firms in Eastern Europe experience as they joined the EU in 2004 was already being experienced by the firms during the accession period, and must be immediately observable in this analysis from 2004 when the Eastern European states joint the EU (Vliegenthart & Horn, 2007; Grabbe, 2001). Nivorozhkihn (2005) argues that Eastern European firms were already adjusting their speed at a similar pace as their western European counterparts. This research was conducted a year after Eastern Europe joint the EU (Joeveer, 2013).
14 incoherent and ever-changing business regulations (the consistently changing tax codes in eastern European countries makes it difficult for firms in this region to calculate their taxes at a cost-efficient manner). Also, the financial system of developing countries in transition to developed countries still demonstrates favorable credit conditions to large firms as banks lack the willingness to finance small firms. This is because Small firms in developing countries lack the collateral or the liquidity constraints required to secure external financing. These firms resort to obtaining credit informally by bribery to secure access to capital. Generally, entrepreneurs in transition economies seek to circumvent bureaucracy required by financial institutions in obtaining loans. In doing so, they obtain loans illegally in the hope of reducing transaction costs (Joeveer, 2013; Joeveer, 2006; Hellman, Jones, Kaufman, & MarkSchankerman, 2000; Aidas & Adachi, 2007; Guseva, 2007; Guseva, 2007; Tonoyan, Strohmeyer, Habib, & Perlitz, 2010; Berger, 2007). Moreover, developed countries have a stable legal framework and well-protected property rights that promote planning, resource acquisition, and coordination. This prevents the expropriation of the sources of external financing for firms. Since joining the EU in 2004, Eastern European firms still experience the incapability of their legal, political and financial system in enforcing property rights and legal decisions, as well as resolving business disputes legally, similar to their Western European counterparts. Eastern Europe has adopted the political, legal, and financial institutional systems of western Europe but has proven inefficient in implementation. As a result, firms usually resort to informal networks to compensate for the failure of the institutional systems. Through bribery of public officials, firms bend the institutional rules. This background is evidence that Eastern Europe is still characterized by the lower efficiency of the legal political and financial institutions, and the lack of enforcement. These institutional factors determine the firms’ decision making by signaling the capital choices that these firms make, determine the leverage levels that are most acceptable and supportable given the risks that the institutions provide. These Eastern European institutional characteristics explain the reason for the volatility in this region. In volatile markets, firms may deviate from their desired leverage levels because of the frequent capital cost adjustments. These arguments provide reasons why Eastern European firms capital structure differ from the capital structure of Western European firms (Hellman, Jones, Kaufman, & MarkSchankerman, 2000; Aidas & Adachi, 2007; Guseva, 2007; Guseva, 2007; Tonoyan, Strohmeyer, Habib, & Perlitz, 2010; Vliegenthart & Horn, 2007; Parkert, 2007; Berger, 2007; Joeveer, 2013; Joeveer, 2006).
2.2.2: hypothesis Development
15 Demirguc-kunt and Maksimovic (1999) examine how differences in financial and legal institutions in 30 countries in the period 1980-1991 affect firms’ use of debt and choice of debt maturity. Their study is from the perspective of developed countries versus developing (emerging market) countries. They find that in inefficient and costly legal systems that characterize developing country markets, investors reduce the risk of financing by offering lower short term debts to firms, while shareholders find no financial relevance to investing (Demirguc-kunt and Maksimovic, 1999).
Cai, Tan, Wang, and Zhong (2019) investigate how Chinese regional political influence favor firms located in regions with political advantage. They find that firms with better institutional favoritism receive more loans with longer maturity (long-term) debt from banks. According to Connor (2009), being part of the EU does not mean that all political risks in Eastern Europe have been eliminated. The paper argues that political risk is still in existence, but remains subtle in Eastern Europe, and can result in delays in the creation of new firms and financing of existing or new firms.
By analyzing the cause of differences in foreign bank ownership between new Europe (Eastern Europe) and old Europe (Western Europe), Berger (2007) find a 70% Western European bank ownership penetrating Eastern Europe’s financial systems. This is higher than the 15% Eastern European bank ownership in Western Europe’s financial system. The reason for such high Western European bank penetration in Eastern Europe is relaxation in government barriers for foreign competition. Relaxation of such barrier results from the influence of the EU in the political infrastructure in Eastern Europe. The penetration of Western European banks into Eastern Europe changed the financial system of the geographic environment, exposing firms in this region to more flexible financial resources and better access to external debt financing (capital). The presence of more open financial intermediaries improves efficiency in collecting firms’ information, giving investors more incentive to provide financing. Firms’ ability to obtain debt capital improved, resulting in improved leverage (Pietro, Palacin-Sanchez, & Roldan, 2018) (Berger, 2007).
16 is developed influence the firm’s access to external capital. An increase in the level of country financial development increases the level of growth opportunities for firms in the country. An increase in growth opportunity reduces financial constraints attached to firms and reduces the cost of obtaining financing (Demirguc-Kunt & Maksimovic, 1999; Love, 2003; Rajan & Zingales, 1995; Gonenc & Haan, 2014 (Khurana, Martin, & Pereira, 2006)).
The resulting conditions leading up to the access to financing is different across economic markets and industry (Pietro, Palacin-Sanchez, & Roldan, 2018). By being part of the EU, Eastern European firms will benefit from and have equal access to the harmonized legislative and regulatory system of the EU. However, legal risk is still a factor in this region because of poorly trained judiciaries and uncertainty in the application of the law and regulatory regimes (Fernandes, 2011; Connor, 2001). Thus, Firms receive less equity and debt financing because lenders and shareholders are reluctant to finance possible positive net present value projects. Lenders’ and shareholders’ reluctance arise because of their fear of poor legal protection on their investments from the legal system. According to (Khurana, Martin, & Pereira (2006), costly external financing limits firms’ ability to obtain external financing, forcing firms to rely solely on internal financing. Similar to Gonenc and Haan (2014), this paper expects the level of institutional development to increase the level of leverage in developing countries. This argument suggests the third hypothesis as follows:
𝐻30: Firm internationalization and the level of country development do not result in a difference in the capital structure of Eastern European firms and Western European firms.
𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑊𝐸 = 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐸𝐸
𝐻31: Firm internationalization and the level of country development result in Western European firms significantly using more leverage in their capital structure than Eastern European firms
𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑊𝐸 > 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐸𝐸
Where 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑊𝐸 proxy the capital Wester European Firms and 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐸𝐸 proxy the capital structure of Eastern European firms.
2.3. The global financial Crisis
17 that depend on these banks for financing. They find that firms that depend on banks for their financing lost disproportionately higher market value and suffer larger declines in their capital investments and profitability during the crisis. Lemmon and Roberts (2010) also looked at how the capital structure is affected in a crisis by studying how shocks in the supply of credit impact corporate financing and investment based on crisis events in 1989. The paper finds that shifts in the supply of capital have consequences on the financial and investment policies of firms because of a decline in their debt and equity investment. This has offsetting effects in the corporate leverage ratios. Cheva & Purnanandam (2011) and Lemmon & Roberts (2010) conclude that these shocks reduce firms’ capacity to gain access to debt and equity financing. Cheva & Purnanandam (2011) further argues that the global integration of financial systems allows financial shocks to bleed across economies through the banking sector. This explains how a subprime crisis in the U.S. extended to affects Europe and the rest of the world. Other studies carried out from different perspectives find mixed results. Hoang, Gurau, Lahiani, and Seran (2018) study the effect of the global financial crisis on the effect of total debt, long-term debt, and short-term debt of French Micro-enterprises. The paper finds that during the crisis, the micro-enterprises increase their use of short-term debt to improve operational growth but reduce the use of long-term debt investments on intangible assets. These pieces of evidence prove that the global financial crisis can result in serious economic consequences (Pindado, Requejo, & Rivera, 2020). To counteract the effect of a financial crisis, central banks buy private and public debt of firms through their monetary policies to improve liquidity (increase credit and money supply). An increase in the supply of money in the economy improves firms borrowing due to higher levels of liquidity in the market. Therefore changing the amount of money in the money market is an effective tool in periods when there is a financial crisis (Pindado, Requejo, & Rivera, 2020). Kantor and Holdsworth (2010), who argue that the main regulation in times of financial crisis is for the central bank to add liquidity in the economy without any limit, echo this sentiment. Kantor and Holdsworth (2010) further agree with other studies that during a financial crisis, Shareholders are reluctant to provide equity capital while debt holders are reluctant to provide debt financing. These studies provide contradictory results. A Financial crisis decreases the use of leverage since the reduction in credit supply as a result of the global financial crisis is expected to reduce firms’ use of external capital. However, by introducing liquidity into the lending system, firms can maintain their leverage in the event of a financial crisis. During a crisis, firms also use up their internal resources, which may force them to depend on external financing for survival (Hoang, Gurau, Lahiani, & Seran, 2018). Based on these arguments, this paper posits that it is inconclusive whether the capital structure of European firms will be positively affected or negatively affected in the aftermath of the global financial crisis. Thus, the fourth hypothesis is that:
𝐻41: The 2007-2009 global financial crisis does not significantly affect the capital structure of Eastern European firms and Western European firms.
18 𝐻41: The 2007-2009 global financial crisis significantly affect the capital structure of Eastern European firms and Western European firms.
𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑃𝑅𝐸 ≠ 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑃𝑂𝑆𝑇
19
3. methodology and Data
3.1. Methodology
Existing literature shows that both firm-specific variables and country-specific variables affect the capital structure of firms (Rajan & Zingales, 1995; Gonenc & Haan, 2014; Pietro, Palacin-Sanchez, & Roldan, 2018). This paper simultaneously uses both firm-specific variables and country-specific variables for this analysis. On the one hand, foreign performance and foreign asset ownership variables are used to give the idea that this analysis is tailored to the specifics of firm internationalization. Additional variables (liquidity and other performance variables) to account for confounding effects. On the other hand, a dummy institutional variable is used to allow for inter-regional economic development comparisons as hypothesized by this paper. The hypothesized dependent variables of interest are firms’ leverage ratios, which are continuous (and not dummy) variables. Following existing literature (Rajan & Zingales, 1995; Gonenc & Haan, 2014), this paper uses linear panel data regression analysis to test the relationship between the leverage ratios and the independent and control variables described next. Existing studies use similar variables in implementing their different research methods. Using variables similar to what other research papers use is advantageous because the variables are independent of the industry or location. In this regard, robust research results with similar standards are obtained irrelevant of which industry or country the firms belong to. Results can be compared across countries and industries for different research findings. Increasing the scope of industry and country implies an increase in the sample size by ensuring data availability for firms from more countries. This is relevant for a reliable analysis of the hypothesis of this paper because a large sample size ensures consistency in the data. By ensuring a sufficiently large sample size, this paper does not require data normality in hypothesis testing. (Ghasemi & Zahediasl, 2012). This paper collects panel data for this analysis, which includes observations on the same cross-sectional firms over time.
3.1.1. Variables
20 These characteristics make the debt to asset ratios suitable to measure and compare the capital structure of firms. The total debt ratio is the ratio between total debt and total assets, the long-term debt ratio is the ratio between long-long-term debt and total assets, and the short-long-term debt ratio it the ratio between short-term debt and total assets. Employing separate use of total, long-term and short-term debt ratios allows this paper to separately analyze the influences on the maturity structure of debt as well as on leverage (Wijst & Thurik, 1993). The fourth proxy for the capital structure that is commonly used in the capital structure literature is the debt to equity (DEquityR) ratio (Pietro, Palacin-Sanchez, & Roldan, 2018; Fernandes, 2011; Mukherjee & Mahakud, 2012; Rajan & Zingales, 1995). The debt to equity ratio accounts for shareholder capital employed into a business and thereby best represents the effects of past investment and financial decisions. This definition of leverage relates to the agency problem associated with debt, as Jensen and Mecklin (1996) and Myers (1977) suggest. A higher debt to equity ratio shows that the firm can meet its financial obligations and that the firm is using debt leverage to increase equity returns for its investors. High debt to equity ratio is positively related to an increase in performance and growth. This paper uses the natural logarithm of each leverage ratios to capture the effect of total debt and firms’ ability to obtain and pay their debt, thereby determining firms’ leverage, (LogTDebtR, LogLDebtR, LogSDebtR and LogDEquityR for total debt ratio, long-term debt ratio, short-term debt ratio and debt to equity ratio respectively). Extant literature on capital structure uses a maximum of three of these leverage ratios. By using all four leverage ratios, this paper emphasizes more on the fact that capital structure theories suppose a different impact for each type of leverage ratio (debt) (Pietro, Palacin-Sanchez, & Roldan, 2018).
Another variable that is very essential for this analysis is internationalization. The parameter needed to determine firm internationalization originates from firms’ sales and are derived from firms’ income statements. This paper uses firms’ international sales to determine firms’ level of internationalization, similar to Mansi and Reeb (2002) and Gonenc and Haan (2014). Dorrenbacker (2002) makes the point that increased globalization is positively related to a quantitative increase in firms' international activities. Dorrenbacker (2002) recommends using foreign sales as a proxy for measuring firm internationalization. Internationalization is measured as the natural log of foreign sales to total sales ratio. Also, the dummy variable for foreign sales (DFSales) is used to compare the leverage of Eastern European multinational firms and Eastern European domestic firms. Using the definition of multination firms as required by the Financial Accounting Standard Board No. 14 (FASB 1976), a multinational firm must report a foreign sales ratio of 10% or higher of total sales. The value 1 is assigned to the dummy variable if it is a multinational firm and zero otherwise.
21 2004 until 2008 just before the crisis’s major effect on European countries in 2009. The second period is between 2009 until 2012 after the financial crisis when Europe was still figuring out how to deal with the sovereign debt situation. The third period is 2013 until 2017. This period determines if the capital structure of the firms recovered to what it was before the crisis hit Europe (Europa, 2020). Two dummy variables are necessary for this analysis. The dummy 2008-2012 is assigned the value 1 if the period is between 2009 until 2012 and the value 0 otherwise. The dummy variable 2013-2018 is assigned the value 1 if the period is between 2013 until 2017 and the value 0 otherwise. Since the regression model has a constant, this paper avoids the dummy variable trap by not including a dummy for the period 2004 until 2008. This paper expects the crisis not to affect firms' leverage ratios because of the role of the central banks to ensure liquidity during the crisis.
23 This thesis controls for country-specific factors through country-level of financial development (FinDev). According to Lin, Kim, and Lee (2014), financial development and foreign direct investment (FDI) are positively related. It is important to control for financial development because it directly leads to economic development (Lin, Kim, & Lee, 2014; Bist, 2018; Love, 2003). This thesis follows the method implemented by Love (2003) to determine financial development, from splitting financial development into stock development and credit development. Standardized indexes for determining stock development and credit development are obtained from the FinancialStructureDatabase. The indices are; liquid Liability over GDP and private sector credit over GDP to measure credit development and market capitalization over the gross domestic product (GDP); total value traded over GDP and total value traded over market capitalization to measure stock development. Aggregating stock development and credit development measure financial development (Khurana, Martin, & Pereira, 2006). Rajan and Zingales (1998) argue that financial development has a substantial supportive influence on the rate of economic growth due to the part it plays in reducing the cost of external financing, increasing the possibility to obtain leverage.
Legal, political, and institutional differences between Eastern European and western European firms constitute what makes countries in these geographic regions differ in their level of development. This paper uses the variable regional institutional and corporate governance (ICG) to measure the difference in leverage between Eastern European and Western European firms. A dummy variable holds the value 1 if the firm is from Western European countries, where market forces are the determinant factors of firms choice of leverage and zero otherwise, where legal, political and financial factors, other than market forces influence the firms’ leverage decisions (Danisman & Demirel, 2019).
This paper uses interaction variables to capture the effect of one variables’ dependency on one or more other variables. LogFSalesR*LogFinDev examines the role played by financial market development in the relationship between firm internationalization and leverage ratios.
LogFsalesR*LogTobinQ examines how growth opportunities and internationalization affect
24 For the robustness test, this paper uses country-level political risk obtained from EIKON to test for model specification (Gonenc & Haan, 2014).
3.1.2. Regression Model
The problem with panel data is that there might be cross-sectional effects on each firm or group of firms. No correlation should exist between the independent variables and the error term, implying that the regression equation is strictly exogenous. This is an essential requirement to obtain unbiased regression estimates. Endogeneity may arise because of omitted variables and measurement errors in some of the independent variables. Omitted variable bias might arise when the regression model does not measure relevant variables such as the managerial ability. According to Matemilola, Noorden, Bgah, & Nassir (2015), unobservable firm-specific factors such as managerial ability affect firms’ decision to obtain capital. Measurement error arises when capital structure decisions depend on the income statement and balance sheet line items that are subject to restatement due to financial regulations and requirements. Profit, assets, and debt of companies are examples of such line items. For example, profitability is measured using total assets. This implies profitability is bound to have measurement error because its estimates can change when financial data is restated (Connolly, 1986). A consequence of measurement error is window dressing whereby, for example, the asset value of the company is overstated resulting in a higher measure of profitability. The reverse is true when asset value is understated resulting in a lower measure of profitability. This paper mitigates the problem of omitted variables extensively specifying the regression model.
25 to determine coefficient estimates. Also, this paper considered to employ a Redundant Fixed Effect test to infer whether unobservable heteroscedasticity is present. The null hypothesis is that the pooled OLS model is the appropriate method to explain the regression coefficient estimates, against the alternative hypothesis that the FE model is the appropriate model to explain the regression coefficient estimates (Wooldridge, 2016; ElBannan, 2017).
This paper does not report the Hausman test results because it is irrelevant for this analysis. This is because this paper is interested in the results of time-invariant dummy variables. Using the FE model does not allow for time-constant independent variables since any variable which does not vary over time will cancel out. By using FE models, this paper loses the ability to determine the influence of the variables that affect the dependent variable but does not change over time. Fixed effect models are designed to study the causes of changes within an entity. Wooldridge (2016), in explaining how to choose between the FE model or RE model, emphasizes that the most obvious first reason to consider in making this choice is if key explanatory variables are constant over time. In this case, the fixed-effect model cannot be employed because it drops the time constant variables from the regression model, which is relevant for this paper’s analysis. However, using the RE in this way requires the assumption that the unobserved effect is uncorrelated with all independent variables. Also, as many time-constant variables as possible should be included among the independent variables. The RE model is suitable for the belief that differences across entities have some influence on the dependent variables.
There must exist a correlation between the error terms for a given entity at different points in time. The drawback of using the RE model is the use of the assumption that the unobservable variables are uncorrelated with the independent variables. Making this assumption disregards the much useful information in different periods. Therefore, the random effect model does not account for serial correlation across time (autocorrelation). To solve the serial autocorrelation problem, this paper quasi demeans the data to generate generalized least squares estimators. This method works well for this analysis because of the large sample collected for this analysis. Thus, the GLS regression methods have asymptotic properties. GLS estimated models work with both balanced and unbalanced data. The benefit if using GLS estimation methods is that it allows for time-invariant independent variables to be estimated. This is because the RE works with the assumption that unobserved effects are uncorrelated with explanatory variables. GLS estimators also account for heteroskedasticity in the errors.
26 produces results with coefficient estimates closer to what the FE model will produce, but far from what the pooled OLS model will produce (Wooldridge, 2016).
This paper implements the use of logarithmic scales in measuring the firm-specific financial variables, and country-specific variables (except for country political risk variable and all the dummy variables) because by using logarithmic scales, the regression results are not always skewed towards large values by suppressing the variation in the variables (heteroscedasticity). Using the natural logarithm brings about normality in the variables and accuracy in measurements. It allows for better scrutiny and detail scaled analysis of data (Gonenc and Haan (2014). The data can be better scrutinized. It also allows the interpretation of result coefficient estimates in terms of economic elasticity, which is better for economic comparison and conclusions.
A fisher test for unit-root tests whether all the variables in the panel data contain a unit root. The fisher test for unit root is used in this analysis because it works for both balanced and unbalanced data. The test result rejected the null hypothesis that all panels contain unit-roots for every variable at the 1% level. This test concludes that the variables are stationary.
If the dependent variable has a causal effect on the independent variable, then endogeneity exists because, in effect, the independent variable is correlated with the (unobservable) error terms. This results in reverse causality. Endogeneity also occurs as a result of omitted variable bias is affecting the time-varying effect in the model. The methods require proses to solve these issues are either time consuming or difficult to implement, and sometimes do not work. (Bellemare, Masaki, and Pepinsky, 2015). This paper does not further address the issues caused by endogeneity.
Based on these discussions, the following GLS regression model analyzes hypothesis one, two, and three, described in section 2:
𝑦𝑖,𝑡− 𝜃𝑦̅𝑖 = 𝛼𝑖(1 − 𝜃) + ∑ 𝛽𝑖(𝑥𝑖,𝑡− 𝜃𝑥̅𝑖) + (𝑢𝑖,𝑡− 𝜃𝑢̅𝑖) (1)
See Appendix A for the description of the dependent and independent variables.
27 To measure the effect of the financial crisis, this paper makes use of time dummy variables, thus time FE models will be suitable for this analysis because it allows the intercept of the regression model to differ over time but not cross-sectionally. A time fixed effect controls for time effects whenever unexpected variation or special events might influence the dependent variable. Time fixed effects capture both observed and unobserved year fixed effects. This paper used the within estimator method for this analysis because the number of entities is large. Therefore, the least square dummy variable method will be impractical to use with many dummy variables in the model. The Hausman test for this model was statistically significant at the 1% level implying that the fixed effect model is preferable over the random effect model.
Thus, the following model will be used to analyze hypothesis 4:
𝑦𝑖,𝑡− 𝑦̅𝑖 = ∑ 𝛽𝑖(𝑥𝑖,𝑡− 𝑥̅𝑡+ (𝑢𝑖,𝑡− 𝑢̅𝑡) (2)
This equation is derived from the FE model by subtracting the time mean of the observations for cross-sectional entities 𝑖. 𝜇𝑖 is the cross-sectional error term with the assumption of zero mean and constant variance, is independent of the individual observation errors (𝑢𝑖,𝑡 ), and independent of the independent variables, 𝑖 is the entity (firm) and 𝑡 is the time. 𝛽𝑖 represent the coefficient estimates of the independent variables 𝑥𝑖,𝑡, 𝑦𝑖,𝑡 represent the dependent variable, where 𝑖 is the entity (firm) and 𝑡 is time in years. The overbar denotes the time averages. With the within estimators, this paper uses ordinary least squares estimations to proceed with the regression.
3.2. Data
28 The initial data come from firms in the Czech Republic, Estonia, Latvia, Lithuania, Slovakia, Slovenia, Poland, Germany, Italy, France, the Netherlands, Belgium, Luxemburg, Ireland, Greece, Spain, Portugal, Austria, United Kingdome (UK), Denmark, Sweden, and Finland. Information from the largest listed firms with available data is collected for each country, up to a maximum of 100 firms per country. This ensures that the final sample contains enough but not an excessive number of firms per country, which helps to avoid regression estimates that are skewed towards specific country and regional country characteristics. EIKON contains firm-level financial data on more than 60000 active and non-active Eastern European and Western European (EU) listed firms (excluding banks). The firm-level data used in this research is collected in the period between 2004 and 2017. Yearly firm-level balance sheet and income statement data related to the firms’ leverage, sales, asset value, equity value, market capitalization, and income are retrieved from EIKON. This data is used to construct the firm-specific variables relevant to this analysis. This paper recognizes that the period 2007 until 2009, which spells the financial crisis may cause swings in firms' capital structure. Therefore, firms’ level of liquidity is used to adjust for the effect of the financial crisis (Pindado, Requejo, & Rivera, 2020).
The banking industry is not included in the sample because the industry is subject to many regulations. Banks also play an important role in determining firms’ capital structure through their unique ability to be liquidity or debt providers. Including this industry will result in self-selection bias. Firm comparability across these different countries is enhanced by making sure that all firms selected for this analysis are listed. Slovakia is eliminated from the sample due to very limited data. Also, firms with limited data are dropped from the sample, resulting in some countries having fewer firms in the sample than other firms. This is peculiar to Eastern European firms. The final sample consists of 21282 firm-year observations from 21 countries, over 14 years, and 1899 firms. the sample is unbalanced.
3.3. Descriptive statistics
29 average short-term debts than their Western European counterparts. For both Eastern European and Western European firms, short-term debts have lower average leverage ratios than long term debts. The overall average sales are lower for Eastern European firms at 12.590 compared to 13.307 for Western European firms. The same is true for the overall total asset average, which is 13.345 for Eastern European firms and 13.794 for Western European firms. This suggests that firms from Western Europe are larger and are more likely to use debt when compared with firms in Eastern Europe. Also, Western European firms record higher average foreign sales and foreign assets than Eastern European firms. Notice that Eastern European firms have on average fewer liquid firms, more firms with foreign holdings, lower financial development, and slower growth than Western European firms. Latvia records the lowest financial development while Luxembourg records the highest financial development.
30 Table 1: Descriptive statistics of the main firm-level and country-level variables by country
Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) Poland 100 -2.284 -2.950 -3.645 4.907 -1.877 -3.131 1.862 -2.176 -6.064 12.395 -7.113 5.213 0.001 0.339 0.000 0.316 0.036 11.938 3.603 Ireland 100 -2.114 -2.354 -4.964 5.044 -0.711 -1.096 1.958 -2.222 -4.789 12.338 -7.082 5.659 0.002 0.651 1.000 0.815 0.369 12.344 1.885 Latvia 100 -2.107 -2.395 -3.061 6.690 -0.964 -2.901 1.246 -1.205 -4.229 10.206 -8.591 4.703 0.000 0.253 0.000 0.142 0.000 9.662 4.034 Hungary 90 -2.100 -2.638 -4.036 7.058 -1.198 -1.617 1.653 -1.774 -3.042 16.842 -9.057 5.322 0.004 0.615 0.000 0.469 0.081 16.340 4.403 Germany 100 -2.054 -2.450 -4.118 5.470 -0.785 -2.226 1.600 -2.179 -4.054 15.760 -7.470 5.921 0.000 0.459 1.000 0.848 0.597 15.346 5.022 Slovenia 34 -2.010 -2.481 -3.085 6.333 -1.358 -1.313 1.120 -1.462 -3.320 11.805 -8.253 5.046 0.009 0.350 0.000 0.527 0.073 11.079 2.628 Luxembourg 100 -1.976 -2.302 -3.925 5.339 -0.427 -0.507 1.931 -1.975 -4.725 13.827 -7.329 6.874 0.000 0.602 1.000 0.641 0.323 12.544 1.640 Czech Republic 33 -1.922 -1.885 -4.713 6.450 -1.011 -1.579 1.780 -1.819 -5.632 16.255 -8.324 5.141 0.000 0.626 0.000 0.623 0.175 15.200 3.473 Denmark 100 -1.918 -2.216 -4.155 5.292 -0.954 -1.530 1.776 -1.991 -3.363 14.131 -7.193 5.829 0.000 0.270 1.000 0.654 0.215 13.731 1.849 Estonia 69 -1.879 -2.204 -4.325 5.135 -0.970 -1.452 2.043 -1.664 -7.455 11.617 -7.103 5.070 0.000 0.331 0.000 0.546 0.075 11.326 2.046 UK 100 -1.828 -1.989 -4.660 5.357 -0.794 -1.201 1.919 -2.114 -5.414 15.481 -7.126 6.151 0.012 0.586 1.000 0.756 0.597 14.943 2.687 Netherlands 100 -1.810 -2.185 -4.190 5.515 -0.969 -1.272 1.727 -2.174 -4.178 13.195 -7.193 6.089 0.001 0.449 1.000 0.773 0.389 12.924 2.388 Belgium 100 -1.704 -2.032 -4.016 5.674 -0.906 -1.502 1.711 -2.100 -3.427 12.936 -7.352 5.682 0.000 0.476 1.000 0.641 0.200 11.826 2.041 Sweden 100 -1.691 -2.086 -3.877 5.346 -0.708 -1.574 2.064 -2.187 -3.628 15.823 -6.919 5.822 0.000 0.264 1.000 0.811 0.397 15.284 1.924 France 100 -1.678 -2.082 -3.936 5.757 -0.810 -1.900 1.513 -2.034 -4.305 15.863 -7.435 5.879 0.000 0.409 1.000 0.865 0.358 15.362 2.665 Spain 100 -1.678 -2.049 -3.284 5.672 -1.080 -1.549 1.650 -1.765 -5.565 13.942 -7.343 6.218 0.000 0.303 1.000 0.730 0.271 13.091 3.553 Finland 100 -1.654 -2.138 -3.882 5.604 -0.809 -1.666 1.847 -2.150 -4.199 12.644 -7.230 5.672 0.000 0.246 1.000 0.794 0.308 12.617 2.022 Austria 96 -1.621 -1.987 -3.676 6.096 -0.796 -1.425 1.542 -1.647 -4.857 12.960 -7.634 5.551 0.005 0.393 1.000 0.796 0.328 12.305 1.826 Italy 100 -1.524 -2.142 -3.587 6.333 -0.963 -1.540 1.193 -2.461 -4.966 14.288 -7.811 5.798 0.000 0.276 1.000 0.566 0.157 13.513 2.694 Greece 99 -1.296 -2.209 -3.554 7.029 -1.385 -2.379 1.202 -1.449 -5.523 11.733 -8.236 5.660 0.001 0.197 1.000 0.475 0.070 11.062 4.900 Portugal 78 -1.253 -1.799 -3.178 6.980 -1.414 -2.620 1.566 -1.769 -6.363 13.391 -8.185 5.794 0.000 0.342 1.000 0.596 0.121 12.713 2.220
31 Table 3 shows the correlations between all the variables of interest. There is evidence that it is very unlikely for multicollinearity to be a concern during this research. A few variables are strongly correlated at the 1% and 0.1% level. Most of the variables are weakly correlated at the 5% level or show no significant correlation at all. The variables either have a positive or negative correlation with each other. This paper observes from Table 3 that profitability has a strong and negative correlation with short-term debt ratio and debt to equity ratio but has no significant correlation with total debt and long-term debt. This suggests that profitable firms seek less short-term debt and increase their use of equity more than their use of debt. It is inconclusive whether profitable firms change their use of total debt and long-term debt. The correlation matrix also shows no relationship between internationalization and total debt, long-term debt, or debt to equity ratios. However, it shows a weakly negative and significant effect on short-term debt. This suggests that internationalization would not result in changes in the use of leverage. Size has a weak and negative significant correlation with growth opportunity. Growth opportunity on the other hand has no significant correlation with total debt, long-term debt, and short-term debt, but has a strong and negative correlation with the debt to equity ratio. This suggests that an increase in growth opportunity has no effect in the debt ratios but results in an increase in equity financing leading to a reduction in the debt to equity ratio.
32 Table 2: Description and comparison analysis of the main variables
Panel A: descriptive statistics # of
Observations
Mean Median max min Standard Deviation (1) LogTDebtR 19237 -1.786 -1.375 6.622 -15.954 1.46 (2) LogLDebtR 17925 -2.200 -1.761 6.622 -12.893 1.585 (3) LogSDebtR 12378 -3.953 -3.511 2.135 -15.192 1.905 (4) LogDEquityR 16564 5.821 5.986 17.583 -8.161 2.081 (5) LogFSalesR 12106 -.912 -.493 6.518 -12.782 1.39 (6) LogFAssetR 2729 -1.578 -1.046 .813 -16.005 1.934 (7) LogRAsset 16990 1.670 1.785 11.215 -4.605 1.070 (8) LogTang 20463 -1.948 -1.411 .472 -14.038 1.758 (9) LogRandD 5893 -4.43 -4.162 .919 -13.109 1.872 (10) LogTAsset 21224 13.727 13.816 22.323 0.000 2.810 (11) LogTobinQ 17999 -7.535 -7.416 3.215 -17.02 1.305 (12) LogFinDev 21242 5.729 5.753 7.139 3.873 .511 (13) DFLiq 18099 .002 0.000 1.000 0.000 .040 (14) DFHolding 16787 .403 0.000 1.000 0.000 .491 (15) DICG 21282 .797 0.000 1.000 0.000 .403 (16) DFSales 16913 .670 1.000 1.000 0.000 .470 (17) DFAssets 9906 .229 0.000 1.000 0.000 .420 (18) LogTSales 20684 13.168 13.315 22.437 0.000 2.965 (19) PolRisk 21282 2.853 2.400 7.500 1.300 1.210
Panel B: comparison of mean and median of variables between Eastern European Multinational Firms and Eastern European domestic Firms
Multinational Firm Domestic Firm Difference
33
Table 3: Pairwise Correlation matrix of Variables
Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (1) LogTDebtR 1 (2) LogLDebtR 0.734*** 1 (3) LogSDebtR 0.300 -0.107 1 (4) LogDEquityR 0.412 0.175 0.483* 1 (5) LogFSalesR -0.000605 0.407 -0.441* -0.327 1 (6) LogFAssetR -0.0323 0.296 -0.476* -0.315 0.675*** 1 (7) LogRAsset -0.281 0.0520 -0.665** -0.712*** 0.342 0.343 1 (8) LogTang 0.0382 -0.0102 0.490* 0.576** -0.276 -0.271 -0.457* 1 (9) LogRandD -0.230 -0.202 0.0524 0.0813 0.279 0.221 -0.195 -0.115 1 (10) LogTAsset -0.00372 0.188 -0.402 0.000525 0.265 0.250 0.270 -0.458* 0.249 1 (11)LogTobinQ -0.0394 0.103 -0.392 -0.923*** 0.355 0.309 0.651** -0.620** -0.155 -0.0177 1 (12) LogFinDev 0.289 0.304 -0.175 -0.335 0.488* 0.502* 0.296 -0.506* 0.0616 0.372 0.481* 1 (13) DFLiq -0.164 -0.0683 -0.112 0.0419 -0.0572 0.277 -0.0886 0.135 0.112 0.0725 -0.102 -0.0577 1 (14) DFHolding -0.434* -0.0277 -0.674*** -0.138 0.331 0.490* 0.390 -0.240 0.0835 0.451* -0.0460 0.219 0.323 1 (15) DICG 0.574** 0.509* -0.110 -0.224 0.484* 0.297 0.103 -0.490* 0.136 0.182 0.499* 0.791*** -0.129 -0.0780 1 (16) DFSales 0.264 0.482* -0.471* -0.419 0.594** 0.553** 0.390 -0.550** 0.154 0.488* 0.565** 0.608** 0.0517 0.261 0.704*** 1 (17) DFAssets 0.00771 0.287 -0.504* -0.497* 0.649** 0.427 0.405 -0.557** 0.183 0.495* 0.564** 0.646** 0.228 0.366 0.648** 0.832*** 1 (18) LogTSales -0.0372 0.143 -0.457* -0.0541 0.263 0.223 0.315 -0.499* 0.255 0.984*** 0.0324 0.334 0.0829 0.426 0.188 0.529* 0.535* 1 (19) PolRisk -0.151 -0.399 0.201 0.451* -0.424 -0.569** -0.443* 0.309 -0.00299 0.125 -0.550** -0.282 -0.0317 -0.0635 -0.327 -0.398 -0.174 0.121 1
34
4. Regression Results
4.1. Regional development, Firm internationalization, and capital structure
Table 4 reports the RE GLS regression model estimates on the role that regional institutional practice plays in the relationship between firm internationalization, Financial development, and the four levels of capital structure (total debt, long-term debt, short-term debt, and debt to equity). The regression is run on the entire sample of Eastern European firms and Western European firms. For each leverage ratio, two models are estimated. Model (1), (3), (5), and (7) are estimated without considering institutional characteristics (restricted models), and models (2), (4), (6), and (8) are estimated considering the institutional characteristics (Unrestricted model), distinguishing between Eastern European firms and Western European firms by using the dummy (DICG) for Western European institutional factors (developed countries).
All the eight models show identical predictions in sign and statistical significance in almost all cases. The results from all the models show positive and negative coefficient estimates between internationalization (LogFSalesR) and the leverage ratios, but the coefficient estimates are statistically insignificant. This paper fails to reject the null of the first hypothesis. This result matches the result suggested by the correlation matrix. This finding is in line with (Gonenc & Haan, 2014) who find that firm internationalization results in no significant change in the capital structure of firms, but contradict (Reeb, Mansi, & Allee, 2001) who find that firm internationalization is associated with a lower cost of debt financing, and this, in turn, increases the use of debt for those firms. The hypothesis of this paper predicted similar results as (Reeb, Mansi, & Allee, 2001)
35 Tabel 4: Internationalisation, Financial development, and capital structure.
(1) (2) (3) (4) (5) (6) (7) (8) VARIABLES LogTDebtR LogTDebtR LogLDebtR LogLDebtR LogSDebtR LogSDebtR LogDEquityR LogDEquityR
LogFSalesR 0.184 0.174 0.036 -0.031 -0.246 -0.042 0.159 0.162 (0.168) (0.177) (0.222) (0.234) (0.368) (0.401) (0.199) (0.209) LogRAsset -0.048*** -0.047*** -0.023 -0.022 -0.098*** -0.096*** -0.096*** -0.095*** (0.015) (0.015) (0.020) (0.020) (0.033) (0.033) (0.017) (0.017) LogTang 0.232*** 0.246*** 0.265*** 0.284*** 0.154*** 0.160*** 0.238*** 0.255*** (0.020) (0.020) (0.026) (0.026) (0.035) (0.035) (0.024) (0.024) LogTAsset 0.130*** 0.118*** 0.179*** 0.167*** 0.001 -0.005 0.160*** 0.145*** (0.013) (0.013) (0.017) (0.016) (0.020) (0.020) (0.016) (0.015) LogTobinQ -0.102*** -0.114*** -0.131*** -0.147*** -0.092*** -0.109*** -0.152*** -0.170*** (0.014) (0.014) (0.019) (0.019) (0.030) (0.031) (0.017) (0.017) DFHolding -0.003 -0.002 0.021 0.021 -0.028 -0.028 0.009 0.010 (0.021) (0.021) (0.028) (0.028) (0.047) (0.047) (0.025) (0.025) DFLiq -0.126*** -0.127*** 0.163*** 0.161*** -0.414*** -0.415*** -0.244*** -0.246*** (0.027) (0.026) (0.035) (0.035) (0.058) (0.058) (0.031) (0.031) DICG 0.886*** 1.087*** 0.468** 1.225*** (0.129) (0.159) (0.186) (0.151) LogFinDev 0.058* 0.012 -0.075* -0.132*** -0.295*** -0.368*** 0.137*** 0.071* (0.032) (0.034) (0.043) (0.045) (0.070) (0.076) (0.038) (0.040) LogFSalesRLogFinDev -0.054** -0.072** -0.034 -0.044 0.050 -0.023 -0.049* -0.080** (0.025) (0.031) (0.033) (0.041) (0.055) (0.070) (0.030) (0.036) LogFSalesRLogTobinQ -0.007 -0.013 -0.010 -0.016 0.000 -0.009 -0.007 -0.015 (0.008) (0.008) (0.010) (0.011) (0.017) (0.017) (0.009) (0.010) LogFSalesRLogRAssetR 0.046*** 0.045*** 0.059*** 0.058*** 0.035 0.034 0.037*** 0.036*** (0.012) (0.012) (0.016) (0.016) (0.026) (0.026) (0.014) (0.014) LogFSalesRDICG 0.086** 0.088 0.182** 0.132*** (0.043) (0.056) (0.087) (0.050) Constant -4.200*** -4.620*** -5.057*** -5.620*** -2.168*** -2.189*** -4.367*** -4.955*** (0.306) (0.311) (0.394) (0.401) (0.564) (0.569) (0.361) (0.367) Observations 7,901 7,901 7,682 7,682 6,482 6,482 7,846 7,846 Number of Firm 861 861 845 845 787 787 859 859 Chow test 47.98*** 48.34*** 7.26** 66.77***
36 Vliegenghart and Horn (2007), Eastern Europe has not yet reached the level of country development of their Western European counterparts despite both geographical regions sharing the same European economic market environment.
This paper controls for the interaction effect between internationalization and the dummy for Western European firms (LogFSalesRDICG). The results are positive and statistically significant at the 5% level for total debt and short-term debt, at the 1% level for debt to equity, and show no statistical significance for short-term debt. This indicates that ceteris paribus, Western European firms that engage in international sales use on average more total debt, short-term debt and debt to equity than their western European counterparts do. This finding contradicts Fan, Titman, and Twite (2012) who find that firms in developed market environments, with legal capital market protection for claimants, use more equity financing and less debt financing as the results in this paper depict.
37 Roldan (2018) who argue that size has a positive effect on firms' leverage. A possible explanation is that debt capacity increases with an increase in company size. However in large companies that show limited growth opportunity, size is negatively related to leverage ratio ( (Sari, Siska, & Sulastri, 2019).
The results for growth opportunity (LogTobinQ) is statistically significant at the one percent level and has negative coefficient estimates with all the leverage ratios. This result is in line with Gonenc and Haan (2014) who find a negative relationship between growth opportunity and leverage ratios. The theory of information asymmetry (Myers, 1984; Myers & Majluf, 1984) can explain that growth firms tend to be more difficult to value. However, Chen (2004) finds a contrasting result that growth opportunity is positively related to firm leverage because, by signaling growth, these companies expropriate financing from eternal investors.
Tangibility (LogTang) shows a positive and significant relation with all the leverage ratios at the 1% significance level. This result is in line with the existing literature. Chen (2004) argues that this positive relation exists because tangible assets increase the collateral security for lenders and therefore reduce the risk of external financing. Jensen and Meckling (1976) also make the same argument for a positive relationship between tangible assets and external financing. Gonenc and Haan (2014) also find a similar positive relation between tangibility and leverage.
Tables 4 shows mixed results between Liquid and illiquid firms. Liquid firms have a negative relation with total debt, short-term debt, and debt to equity, but have a positive relation with long-term debt. The coefficient estimates are statistically significant at the 1% level for all the leverage ratios. The results indicate that ceteris paribus, liquid firms have on average 0.127%, 0.415%, and 0.246% less total debt, short-term debt, and debt to equity respectively, but 0.161% more long term debt than non-liquid firms. These findings for total debt, short-term debt, and debt to equity ratio are in line with the modified pecking order theory, which suggests that firms should first use retained earnings, equity, and then long-term debts (Chen, 2004). Implementing this theory implies that managers serve their self-interest and the interest of shareholders by first using firms’ liquid assets for funding (internal source of funds) instead of using external financing, which results in a lower level of debt (Deesomsak, Paudyal, & Pescetti, 2004). However, a positive relationship exists between liquid firms and long-term debts. More liquid firms prefer to finance operations by using more of internal funds (Saarani & Shahadan, 2013). A possible explanation is that large liquid firms display lower information asymmetry, and this along with high liquidity, can rely on debt financing because of the lower cost of adverse selection (ElBannan, 2017).
