Foreign Ownership, Financial Constraints and Financing Decisions: Evidence from Ghana and China Agyapomaa Gyeke-Dako, MSc. Thesis submitted to the University of Nottingham for the degree of Doctor of Philosophy September, 2010 1 ABSTRACT Firms playa pivotal role in every economy. Therefore their financial standing should be of great concern to policy-makers, as this may directly affect the overall performance of the economy. Corporate finance literature however suggests that market imperfections resulting from conflicts of interest and informational asymmetries, between different economic agents, limit firms in their ability to finance investment projects. Yet, the extent to which firms are limited depends greatly on their size, ownership, exporting status, risk level and even location. In fact, many studies believe that unless by some policy directives pertaining to a particular country, foreign, exporting, large, less risky and firms in developed regions will under normal circumstances be less financially constrained than their counterparts, domestic, non-exporting, small, more risky and firms in undeveloped regions respectively. For domestic firms, some evidence exists that their financial constraints can be alleviated by the entrance of foreign firms. In this thesis, we try and incorporate as many of these determinants as possible bearing in mind data availability. As a result, in the first part of this study, we address the issue of financial constraints of foreign and domestic firms as well as crowding out effects of domestic firms in Ghana. To the best of our knowledge, no study has attempted to address this particular issue on Ghana. We then move a step higher in the second part of our study by considering this same topic in China but then looking at it using firms with different degrees of foreign ownership. Not only do we do this, but we also carry out regional analysis in this respect. All these are made possible by the quality of dataset that we use in our analyses. Again, to the best of our knowledge, no study has addressed this issue on China in this manner. In the last part of our study, we take a different tum altogether and examine mainly the sensitivities of both long-term debt and short-term debt to cash flow and collateral respectively for globally-engaged firms (foreign-owned and exporting firms) as well as firms in the coastal and non-coastal regions of China. We find very interesting results from these analyses. With regards to evidence on firm financial constraints in Ghana, we find that domestic firms 2 are more financially constrained than foreign firms and that foreign firms' presence has no impact on domestic firms in Ghana. For the results on China, while we find no difference between the financial constraints of purely domestic firms and joint ventures, we find that wholly-foreign owned firms are less financially constrained than purely domestic firms. We however find that the presence of foreign firms help alleviate purely domestic firms from their financial constraints. For the regional analysis, estimates based on this suggest firms in the coastal region are less financially constrained than firms in the non-coastal region. Regarding the sensitivities of both short-term debt and long-term debt to cash flow and collateral, whilst we find that highly globally engaged firms have a higher sensitivity of long-term debt to cash flow than non-globally- engaged firms, we find that globally engaged firms' short-term debt have a lower sensitivity to collateral than non-globally engaged firms. As for the regional analysis, our result show that both the short-term debt and long-term debts of firms in the coastal region have a higher sensitivity to collateral and cash flow respectively than the sensitivities of short-term debt and long-term debt to cash flow and collateral of firms in the non-coastal region. 3 ACKNOWLEDGMENT First and foremost, I am very grateful to God for seeing me through this period. Second, I am indebted to my supervisors, Professor Alessandra Guariglia and Dr. Sarah Bridges for their guidance, constructive criticisms and valuable contributions in the preparation of this thesis. I also wish to thank Professor Holger Gorg for playing an important role in this research work. It gives me great pleasure to acknowledge the support of the participants at the 2008 Leverhulme Centre for Globalisation and Economic Policy (GEP) Conference, University of Nottingham for their comments on the third chapter of this thesis. My thanks also go to all the participants at the 20th Chinese Economic Association (UK) Annual Conference held at the University College, Dublin in July, 2009 for their very helpful comments on chapter four of this thesis. To the Leverhulme Centre for Research on Globalisation and Economic Policy, I am most grateful for their financial support. Last but not least, thanks to my family, especially my husband for all their encouragement and support. The usual disclaimer applies. 4 Table of Contents Chapter 1...................................................................................................................... 11 1. Introduction 11 1.1 Background 11 1.2 Motivation and Methodology 16 1.3 Dataset 23 1.4 Summary of Main Findings 24 1.5 Structure ofThesis 25 Chapter 2 27 2. Literature Review on Local Finns, Foreign Finns and Financial Constraints 27 2.1 Introduction 27 2.2 Explaining the Presence of Financial Constraints Using Investment-Cash Flow Sensitivity 27 2.3 Investigating the Existence of Financial Constraints Using Ownership Structure ................................................................................................................................. 28 2.4 Classifying Finns into whether or Not they are Globally- Engaged to Test for Financial Constraints 31 2.5 Examining the Link between Investment and Financial Constraints In Periods of Currency Crises 34 Chapter 3 36 3. Foreign- owned Finns and Financial Constraints: Evidence from the Ghanaian Manufacturing Sector 36 3.1 Introduction 36 3.2 An Overview of Ghana's Financial System and Foreign Direct Investment.. 41 3.2.1 Ghana's Financial Sector 41 3.2.2 FDI Trends and the Macroeconomic Environment in Ghana 46 3.2.3 Ghana's Business Environment.. 50 3.3 Data 52 3.4 Approaches to Testing for the Existence of Financial Constraints and Categorisation Method 53 3.4.1 An Evaluation of the Different Approaches to Testing For the Existence of ............................................................................................................................. 53 Financial Constraints 53 3.4.2 Categorization Method 54 3.4.3. Variable Measurement 56 5 3.5 Model Specification and Estimation Technique 57 3.5.1Model Specification 57 3.5.2 Estimation Technique 61 3.6 Descriptive Statistics in the Form of Mean Values for Debt 64 and Investment 64 3.7 Discussion ofResults 66 3.7.1 Main Estimation Results 66 3.7.2 Robustness Check 68 3.8 Conclusion and Policy Recommendation 69 Chapter 4 75 4 Financial Health and Foreign Ownership: Evidence from Chinese Manufacturing Finns 75 4.1 Introduction 75 4.2 China's Financial System 83 4.2.1 Evolution of China's financial system (1978-2006) 83 4.2.2 Sources of Finance for Enterprises in China 87 4.3 Foreign Direct Investment in China 90 4.3.1 Sources of FDI Inflows in China 90 4.3.2 Trends in FDI Inflows in China 91 4.4 Data and Summary Statistics 92 4.4.1 Data Description 92 4.4.2Summary Statistics 94 0.35 97 4.5. Baseline Specification, Econometric Technique and Results 99 4.5.1 Baseline Specification 99 4.5.2 Econometric Technique 101 4.6Estimation Results and Robustness Checks 103 4.6.1 Estimation Results for All Finns 103 4.6.2: Estimation Results for Finns with Different Degrees of Foreign Ownership ........................................................................................................................... 104 4.6.3 Regional Disparities in Firms' Financial Constraints and the Crowding Out Test 106 4.6.4 Estimation Results for Soft Budget Constraints 109 4.6.5 Test for Crowding OU1.. 110 4.7. Conclusion and Policy Recommendations 114 6 Chapter 5 124 5 Global Engagement and Financing Decisions 124 5.1 Theories of Capital Structure and Literature Review 124 5.1.1 Theories of Capital Structure 124 5.2 Earlier Studies on the POT, TOT, Market-Timing and Behavioural Finance Theories 132 5.2.1 Theoretical Work 132 5.2.2 Empirical Work on Capital Structure 147 Chapter 6 170 6 Our Application to China 170 6.1 Introduction 170 6.2 The Transformation of Capital Structure in China's Transitional Economy 178 6.3 Dataset 180 6.4 Variable Measurement, Hypothesis Testing and Summary Statistics 181 6.4.1 Variable Measurement 181 6.4.2 Short-Term Debt versus Long-Term Debt Financing in China 182 6.4.3 Determinants of Capital Structure and Hypothesis Testing 183 6.4.4 Summary Statistics 190 6.5 Model Specification, Methodology, and Econometric Technique 197 6.5.1 Model Specification and Methodology 197 6.5.2 Econometric Technique 199 6.6. Estimation Results 202 6.6.1 GMM Estimation Results for all Firms 202 6.6.2 GMM Estimation Results for Firms with Different Degrees of Foreign Ownership using Short-Term Debt.. 204 6.6.3 GMM Estimation Results for Exporting and Non-Exporting Firm Years using Short- Term Debt Ratio 208 6.6.4 Tobit Estimation Results for Firms with Different Degrees of Foreign Ownership Using Long-Term Debt Ratio 210 6.6.5 Tobit Estimation Results for Exporting and Non-Exporting Firms using Long-Term Debt · · ·..· · ·· 213 6.7.5 GMM Estimation Results for Firms in the Coastal and Non-Coastal Regions using Short-Term Debt 215 6.7.6 Tobit Estimation Results for the Coastal and Non-Coastal Regions Using Long-Term Debt ·..·..·..·..· ·..· 218 7 6.7 Conclusion 220 Chapter 7 230 7 Summary of Findings, Policy Implications, Limitations and Future Work 230 7.1 Summary of Findings 230 7.2 Comparison of Chapters 3 and 4 233 7.3 Comparison of Chapters 4 and 6 235 7.4 Policy Implications 236 7.4 Limitations and Future Work 238 8 References 241 8 List of Tables 3.1 Mean of Foreign and Domestic Firms Expressed in Percentages 65 3.2 Percentages of Foreign and Domestic Firms and Type of Financial Source 65 3.3 Percentage of Each Financing Source by the Different Types of Firm- Years 65 3.4 GMM Estimates for Testing the Degree of Financial Constraints 67 4.1 Mean Values of the Various Variables for Firms in the Manufacturing Sector with Different Degrees of Foreign Ownership 97 4.2 GMM Estimates Using Foreign Capital.. 105 4.3 Soft Budget Constraint Test. 109 4.4 Test for Crowding Out. 110 6.1 Summary of Variables Used and their Expected Signs Under Alternative Theories 190 6.2 Percentage of Different Firms that Export and do not Export in the Entire Economy and in the Different Regions of China 192 6.3 Comparison of the Mean Values of Variables among Different Types of Firms 194 6.4 Mean Values of Exporting and Non-Exporting Firms with their Sample T test with Unequal Variance for Mean Difference 194 6.5 Comparison of the Mean Values of Variables in the Different Regions of China 196 6.6 GMM Estimates for Short-Term Debt of Firms with Different Degrees of Foreign Ownership 205 6.7 GMM Estimates for Short-Term Debt for Exporting and Non-Exporting Firms 208 6.8 Marginal Effects for the Tobit Regression for Firms with Different Degrees of Foreign Ownership 211 6.9 Marginal Effect Estimates for Long-Term Debt for Exporting and Non- Exporting Firms · ···.. ··········· .. ·.. ·· 214 6.10 GMM Estimates for Short-Term Debt for Firms from Different Regions 216 9 6.11 Marginal Effect Estimates for Firms in Different Regions 218 3. Al Percentage and Frequency of Domestic, Foreign Small and Large Firms 71 3. A2 Percentage of Foreign and Domestic Firms Found In Each Industry 71 3. A3 Results for Small and Large Firms as well as OLS Results 72 4. Al GMM Estimation Using Single Variables .118 4. A2 GMM Estimation Results Using a different set of Classification 119 4. A3 GMM Estimation Results Using Registration 120 4. A4 Soft Budget Constraint Test Using Registration 121 4. A5 Crowding Out Test Using Registration .122 4. A6 Estimation Results Using Provincial Dummies to Construct Share of Foreign Debt. 123 List of Figures 3. Al GDP Growth Rate (%) (1991-2008) 73 3. A2 Share ofIndustry and Subsectors in Real GDP (%) (1991-2008) 73 3. A3 Growth Rates of Industry and Subsectors (%) (1991 2008) 74 3. A4 Allocation of Credit by Sector (%) (1991-2007) 74 10 Chapter 1 1. Introduction 1.1 Background Firms, especially manufacturing firms playa pivotal role in every economy as an economy's development is determined mainly by the growth and productivity of its firms. The important role played by manufacturing firms in developing and emerging economies has received a great deal of attention over the past years. Not only are manufacturing firms important for the proper functioning of the business sector, but they also provide employment and promote growth in the economy. In order to play their role effectively in the future, there is the need for researchers and policy makers to identify this role and constantly revise their financial positions to develop the best financial and economic environment to harness these benefits. As a result, how they finance themselves should be of prime focus. More often, firms incur substantial up-front costs which they can profitably recoup after realizing sales revenues. These may either involve costs that may be paid once upon entry into the industry (like sunk cost) or recurrent fixed per-period costs (for example, advertising costs). Unfortunately, firms are not always able to meet these financial needs which make the issue of finance central to any investment decision. Extant literature has contended that a major obstacle to firms' investment is financial constraint and the situation is even worse for firms in small open economies (for example, Aghion et. aI.2004). This means that retained earnings or cash flow are often not enough to meet the financial needs of firms and they routinely have to rely on external financing for their operations. External financing may either come in the form of bank loans, trade credits or equity. Yet, Banerjee and Duflo (2005)'s survey evidence on many less developed countries shows that borrowing interest rates are often too high (60% or above) even though deposit rates are less than half as much, and defaults are rare. This puts a great constraint on the borrowing capacity of firms. Why exactly should credit constraints matter for firm development? As well known, credit constraints of firms affect them directly 11 by preventing their expansion and subsequently growth. However, the effect of credit constraints goes beyond capital markets and the firm. Acemoglu (200 I) for instance provides empirical evidence on Western Europe to argue that credit constraints may contribute greatly to unemployment as they limit the extent to which new firms can be created. This problem of credit constraint can be solved if we were operating in the world suggested by Modigliani and Miller (1958). Modigliani and Miller (1958) argued that in the presence of perfect information, firms could easily substitute external capital with internal capital to finance their investment activities, because there was no difference between these two. Consequently, changes in internal capital will not affect firms' investment decision, leaving all firms financially healthy. Following Modigliani and Miller's (1958) seminal paper, economic theory has persistently shown that while the situation in a perfect informational world may serve as an important benchmark in theory, in the real world, capital market imperfections may exist, due to asymmetric information. This has led many scholars to believe that firms may be heterogeneous in their financial constraints. As a result, models of heterogeneous firms have, in the last few years, been at the hub of most theoretical and empirical developments in international finance. Theoretical arguments concerning agency costs fall in line with why firms may be heterogeneous in their degree of financial constraints. Underlying these arguments is the idea that, in the presence of capital market imperfections, asymmetric information about borrowers' type and the likelihood of debt repayment creates a wedge between the interest rate on borrowing and the lending rate, with the former being higher. This translates into an "external finance premium" for firms that borrow, that is, external finance becomes relatively more expensive than internal finance (Bernanke and Gertler, 1995). The external finance premium is likely to be higher for some categories of firms; small, young, risky and domestic firms (Gertler andGilchrist, 1994); Guariglia and Mateut ,2010). Accordingly, due to the higher external premium for these firms, their investment levels are more likely to be affected when there is a change in their internal finance as they cannot easily substitute external finance for internal finance. 12 Fazzari, Hubbard and Petersen (FHP, hereafter) (1998) were the first to introduce heterogeneity in the empirical analysis on the link between investment and financing constraints. They interpret higher sensitivities of investment to cash flow as evidence of higher financial constraints faced by firms. The idea behind this interpretation is that if firms find it too expensive to access external finance, then their investment activities will depend greatly on their amount of internal finance; therefore, a reduction in their internal finance will significantly affect their investment. Several papers have supported this approach (see Bond and Van Reenan, 2002; Hubbard, 1998). However, FHP (1998)'s study did not go without a challenge as some studies have found contrasting results (see Kaplan and Zinagales, 1997,2000; Cleary, 20061 ). More recently, the issue of firm heterogeneity, financial constraints and investment has been extended to firms that are globally engaged. Campa and Shaver (2002) for instance, show that exporters in Spain are less financially constrained than non-exporters. Also, Castaneda (2002) shows that Mexican firms that are export - oriented faced a higher financial constraint than non- exporting firms before the 1995-2000 financial paralysis. Focusing on ownership and financial constraints, Harrison and McMillan (2003), Mickiewicz et.al (2004) and Arbelaez and Echavarria (2002) all find that foreign-owned firms have less binding financial constraints than domestic firms in Ivory Coast, Estonia and Colombia respectively. Guariglia and Mateut (2010) combine the ideas in the previous literatures and show that financial constraints are less binding for globally engaged firms in the U.K (i.e. exporters and foreign owned firms) than firms that are not globally engaged. Not only are foreign firms considered to be less financially constrained than domestic firms but they are also seen to have more debt than domestic firms. There are good reasons for supposing that financial constraints might be lower for globally engaged firms than firms that are not globally engaged and there also very good reasons to suppose that globally-engaged firms may be more leveraged than non-globally engaged firms: Foreign firms find it easier to access the domestic financial market because they enjoy less bankruptcy risk I According to these studies, it is not always true that a monotonic relationship exists between financial constraints and investment-cash flow sensitivities. 13 and adopt international standards faster than their domestic counterparts (Harrison and McMillan, 2003). Moreover, having access to both the domestic and international credit markets can make foreign firms less likely to face financial constraints. Desai et al. (2004b) also point out that foreign firms may receive financial help from their parent company. For exporting firms, because they also depend on demand from foreign countries, they are less prone to the domestic cycle, making them less subject to those financial constraints caused by tight monetary policy and recessions at home/. This provides a more stable cash flow for exporters compared to their non-exporting counterparts, and then makes the exporters more financially healthy (Campa and Shaver, 2002; Garcia-Vega and Guariglia, 2008)3. Finally, a firm can only be an exporter if it is able to meet its sunk cost. Therefore, being an exporter also gives an indication that the firm is productive enough to be able to generate enough profits to recoup the sunk cost that needs to be met when entering the international market for the first time (Robert and Tybout, 1997). This increases the probability that the firm will be able to payoff its external debt, and further reduces the financial constraints that the exporting firm faces. In summary, most of these papers conclude that firms that are globally-engaged are relatively less financially constrained than firms that are not globally- engaged. With regards to the differences in the leverage behaviour between foreign firms and domestic firms, Doukas and Pantzalis (2003) for instance argue that the business and financial risk of foreign firms are usually lower than that of domestic firms because of the diversifications in their operations. Extending their argument, they further state that the foreign exposure foreign firms have allows them to utilise debt financing more as they can use foreign debt as hedging instrument against foreign exchange risk and take advantage of the shortfalls in the macroeconomy of the host country. How then can domestic firms be alleviated from financial constraints? Scholars have assumed that the credit constraints of domestic firms will disappear with the influx of foreign investment. According to Harrison et al. (2004), for instance, foreign investment can be seen as a form of equity finance 2 This argument relies on the assumption that business cycles are not perfectly coordinated across countries. 3 A more stable cash flow gives lenders the assurance that the firm will be able to pay ofTits debts. 14 in which local firms could tap in. In fact some scholars like Aghion et.al. (2004) have also suggested that credit constraints for firms could be eliminated with higher levels of development and this includes opening up the economy to foreign investors. By opening up the economy, foreign investors can bring in scarce capital to meet the needs of the constrained local firms." Foreign Direct Investment (FOI) is increasingly being recognised as a relevant aspect in economic development. More and more countries are striving to create a favourable environment to attract FOI. As a result, attraction of FOI has become an integral part of the policies of many economies, especially in developing economies where capital is greatly limited. In the 1990's for instance, (FOI) became the largest single source of external finance for developing countries (Trevino and Mixon, 2004). In 1997, FOI contributed to about half of all private capital and 40 percent of total capital flows to developing countries (Trevino and Mixon, 2004). In recent times, policy makers, especially m emergmg markets and developing countries have gone beyond creating a more liberal environment, to providing substantial public subventions to influence the location decisions of these investors. These incentives included lowering income taxes or granting income tax holidays, import duty exemptions, and subsidies for infrastructure. While the efficacy of such incentives is often doubted, most countries have still resorted to these incentives as a way of promoting FOI5. The idea behind these incentives often stems from the belief that foreign investment generates positive externalities in the form of technological transfer and they often bring in scarce capital to support domestic firms. Are the actions of these countries then necessarily justified? Some authors (see Fieldstein, 2000 for instance) including anti- globalists have argued that foreign firms will monopolise domestic financial markets and destroy local producers. This fear of being crowded out complicates the structuralisation process in many transitional and developing 4See also Stiglitz (1998), who argues that foreign investment brings additional capital. 5 Both Ghana and China have in recent times directed their policies towards FDI attraction. For instance, as part of its policy, Ghana was able to increase its net foreign direct inflows from 14 million dollars in 1990 to about 106.5 million dollars in 1995 and further to about 144.97 million dollars in 2000. By 2000, it had increased its FDI inflows to about 970 million dollars (WDl, various years). Also in China, prior to 2008 they were also given exemption from corporation tax for two years after which they were allowed to pay taxes at a reduced rate of50%5. In addition to the above benefit, they enjoy duty free imports of components and supplies (Naughton, 2007). 15 countries. These scholars further argue that even though spillovers exist, more efficient foreign firms may draw demand from the domestic financial markets making domestic firms less attractive, so the negative competitive effect may outweigh the positive spillovers. Other studies (Caves, 1996; Blomstrom et aI., 2000) argue that the likelihood that foreign firms will "crowd out" local companies is larger in developing than in developed countries, because of a higher technological gap that exists between domestic and foreign firms. These results often raise doubts about the significance of FDls and whether FDIs incentives are indeed justified. The role of incentives in promoting FDI has increasingly been the focus of many studies but their relative advantages and disadvantages have never been clearly understood. Indeed, as much as some spectacular successes have accompanied the attraction of FDIs in some cases, in other cases, there have been some failures associated with providing incentives to attract FDIs. At the end of this study, we believe that policy makers of both countries will find this study a helpful tool in the design, implementation and administration of such incentives. Not only are we concerned about the impact of FDI on domestic firms on the host countries, but this may also provide important implications for the managers of FDIs as it affects their bargaining position and reputation in host countries. 1.2 Motivation and Methodology To provide more insight into these issues, the study seeks to answer the following questions: I) Are domestic firms more financially constrained than foreign firms? If the answer to this question is yes, then what can domestic firms learn from foreign firms to be able to reduce their financial constraints, 2) is there any evidence for crowding out effect and, 3) are foreign firms' short- term debt and long-term more or less sensitive to cash flow and collateral? Our main contribution is mainly empirical. First, the study uses data to assess the financing behavior of firms and derive critical results for two countries that have not been highly researched into, namely China, a country at an intermediate level of development that has 16 a potential of becoming an instrumental player in the world economy, and a low income country, Ghana, to see whether the results are portable across countries. Second, we observe whether the stylized facts observed from the studies of other countries could also apply to only these markets or whether they had more general applicability. Third, firms in both countries appear to be most susceptible to financial shocks, even though that of China is expected to be milder. Considering that firms are constrained in these countries, negative shocks to cash flow can curb their investment activities. We do these in the first and second parts of our study, that is, in Chapters 3 and 4 of this thesis. Since we are conducting our study on countries with different institutional, legal and financial environment, our study affords us the opportunity to see if these differences account for the differences in results. We therefore proceed to compare the results we have for Ghana and China in our main conclusion. Whilst the literature has mainly focused on the role of state ownership of firms (see for example, Dewenter and Maletesta, 2001) and of banks (see for example, Sapienza, 2004), many firms in developing and emerging economies are now either privately owned, joint ventures or foreign investors. These different ownership structures may bring about different governance mechanisms within the firms. Based on this argument therefore, research on foreign firms is warranted. In a part of this study, we will build a model of investment and financing that tries to incorporate heterogeneous firms. In all of these studies we adopt an approach similar to that of McMillan and Harrison (2004) but we add several distinctly new contributions. Like McMillan and Harrison (2003), we estimate a structural model based on the Euler equation for investment and use the Generalised Method of Moments (GMM) to examine the extent to which firms are financially constrained and whether incoming FDI alleviates domestic firms' financial constraints. For our analysis on Ghana, we classify firms into whether they are domestic or foreign. Different from McMillan and Harrison (2003), most of our foreign firms represent individual foreign firms whose owners are foreigners but have lived most of their lives in Ghana. This therefore means that the types of foreign firms we consider are not necessarily FDls. We provide further insights on these categories of firms in the next chapter. To the best of our knowledge, no study has actually looked at foreign firms of this 17 caliber. It is our belief that results on this will further expand our understanding on foreign ownership and their operations. However, for our analysis on China, we provide fresh evidence on firms with different degrees of foreign ownership by including a third, category, Joint Ventures. The reasons why we do this is that, first, the dataset makes this segregation possible. Second, including Joint Ventures will bring out the main relationship between foreign ownership and financial constraints, and third, being a Joint Venture in China is likely to place you in an advantageous position as both the foreign investor and the Chinese partner are likely to benefit: On one hand, the foreign investor is likely to benefit from the government support, land and access to suppliers while on the other hand, domestic firms could take advantage of knowledge spillovers, and access to foreign capital. This could mean that Joint Ventures are likely to face less financial constraints. As a result classifying part of Joint Ventures as domestic firms and part of them as foreign firms may lead to serious aggregation problems as the results of Joint Ventures may not be particularly singled out. Prior studies on financial constraints have neglected this fact. Bridging this gap then becomes our second contribution to literature. In both the studies on Ghana and on China, we further investigate whether foreign firms' presence are able to crowd out domestic firms. After carrying these analyses separately, we then proceed to compare the results of China to that of Ghana, but this is done at the latter stage of this study, that is, in the conclusion. This brings us to the third contribution of our study to literature. Previous studies have concentrated on individual countries but have not provided any comparative studies. Our aim is to find out where firms in a developing economy stand in terms of financial constraints and crowding out compared to firms in an emerging economy. We also seek to find out if there are any lessons Ghana could learn from China to be able to converge to the level of economic performance China has at this point in time. If indeed we find there are any lessons to be learnt from China, our aim will be to suggest some policy implications for Ghana. There are various reasons why we base our studies on China and Ghana. First, for China, after more than 20 years of economic reforms, the number of state-owned enterprises is gradually receding in the economy and 18 being eclipsed by foreign firms and firms with private ownership (Li.et al. (2009». China is now among one of the top destinations for FDI in the world. It has accounted for about a third of gross FDI flows to all emerging countries and about 60% of these flows to Asian emerging economies (Prasad and Wei, 2005). The rise in the number of foreign firms has to some extent changed the Chinese corporate setting from a much localized setting into a more global setting. Second, China is among the countries with the highest FDI inflows in the world, so there should be a sufficient foreign presence to address my questions. Thirdly, both wholly-foreign and joint ventures now exist together in China, which makes it easier and particularly important for such a study to be carried out. Until about a decade ago, wholly-foreign owned firms were not allowed to enter the Chinese market. They had to invest mainly in large and well-functioning state-owned enterprises in the form of joint ventures. Fourth, China, until 2008, had various policies in place which encouraged foreign inflows. This included a total eradication of their corporate tax in the first three years of their establishment and a 50% reduction in the subsequent years. Our objective is to find out whether the policies were warranted as our study falls within this period. Fifth, turning to the financial system of China, it is dominated by a large underdeveloped banking system mainly controlled by the four largest state-owned banks. The government still directs the banks to lend to the state-owned enterprises and provide lending rates and reference borrowing, paying less attention to financial considerations. This makes the financial system an interesting case to study. Finally, being a model for countries that started transition later, the results might be generalised to other countries in transition. For Ghana, despite the manufacturing firms' importance, no study has actually looked at their financial constraints based on foreign ownership, let alone the foreign firms of our caliber. Second, in recent years (that is, from 2001-present), there has been a tremendous increase in the number of banks and other financial institutions in Ghana. In about 1995, the number of banks was only seven. This number rose to about fourteen in 2000 and twenty in 20056, and by 2007, the Bank of Ghana had licensed about twenty three banks 6The evidence of number of banks was taken from a survey conducted by Price water house coopers in 2010. 19 to operate in the economy. Our aim is to find out if the increase in the banking density has had any impact on the availability of credits to firms, especially domestic firms who have been well-known in other countries to face financial constraints. Thirdly, Ghana has enjoyed tremendous growth in recent times. From an average growth rate of about 4.6% in the 1990s and hovering around that same figure in some part of 2000, its growth rate surged to about 7.3% in 2008. Given this increase, our aim is to find out if the increase in growth could be attributed to expansion of credit in the immediate preceding periods before 2008 to firms in the economy, especially domestic firms. Third, the government of Ghana has had widespread economic liberalization in a bid to attract FDI inflows and solve the problem of economic stagnation and decline. This has led to a tremendous increase in the amount of FDI in Ghana. For example FDI net inflows increased from about 67.8 million dollars in 1970 to about 970 million dollars in 2007. Our main concern is whether Ghana is justified in attracting such a tremendous amount. Our fourth contribution to literature is that we examine the capital structure of firms with different degrees of foreign ownership in China. The question we mainly address is, "do firms with a higher degree of foreign ownership have their leverage ratios being more sensitive to cash flow and collateral than firms with a lower degree of foreign ownership? In doing this, we also try to incorporate as many determinants of capital structure as we can in our model bearing in mind data availability. We set up some hypotheses for this in Chapter 6 to explain how this should work and test these hypotheses. With this particular section, we divide firms into three groups: These include Purely Domestic Firms, Joint Ventures and Wholly-Foreign Firms. Like the study on firm financing constraints in China described above, we have included Joint Ventures in this study to single out their possible leverage behaviour. As far as we are concerned, Li et a1. (2009) is the only study that has considered foreign ownership in the literature of capital structure on China. Yet, their study involves, 1) a different category of firms, that is, foreign-owned, private and state-owned firms, and, 2) they just investigate the capital structure of the individual firms and neglect the possibility that the determinants may vary across the different groups. Explaining the hypotheses set in this particular chapter in brief, for the same reasons highlighted above on why foreign firms 20 could a priori be considered as less financially constrained than domestic firms, many theoretical arguments support the view that foreign firms' debt ratios should be different from that of domestic firms (see Doukas and Pantzalis, 2003 and Desai et al. 2004). We use both long-term debt and short-term debt ratios as our leverage ratio measures. With regards to the sensitivities of short- term and long-term debt to cash flow, whilst we expect the sensitivity of short- term debt to cash flow to be lower for Wholly-Foreign firms than Purely Domestic firms, we expect the sensitivity of long-term debt to cash flow to be higher for WF firms than Purely Domestic firms. For Joint Ventures, they may swing either way. To summarise briefly our reasons, because most foreign firms are usually affiliated internationally, it is often argued that their activities allow them to create extensive internal markets that are likely to provide them with cheaper financing than the external markets (Doukas and Pantzalis, 2003). This means that if internal markets function properly, one would expect foreign firms to depend more on internal financing than external financing. However, this argument in our opinion will apply to short-term debt which is usually used to finance short-term projects according to the maturity matching principles. This is because with short-term projects being relatively cheaper than long-term projects, the large pool of internal financing readily available to foreign firms should be enough to cater for their short-term projects. Bearing this in mind therefore, any increase in cash flow should reduce the amount of short-term debt drastically for WF firms, making the sensitivity of short-term debt to cash flow lower for WF firms than that of purely domestic firms. For long-term projects which are supposed to be financed by long-term debt, also according to the maturity matching principle, we do not expect internal financing to be sufficient to cover this. This is when the efficiency and performance of firms can count towards receiving external financial assistance. In this regard, WF firms are in a better position to be able to secure more external financing because of their better credibility. Several authors (Lewellen, 1971; Williamson, 1975, 1986; Myers and Majluf, 1984; Shleifer 7 It is worth-mentioning that this particular study is not just on foreign firms but specifically related to multinationals. Therefore, as a matter of cautioning, although all multinationals are foreign firms, not all foreign firms are multinationals. 8 According to this principle, short-term debts are used to finance short-term assets whilst long-term debts are used to finance long-term assets. For more insight about the maturity matching principle see Myers (1977). 21 and Vishny, 1992; Stein, 1997) provide support to this by stating that diversified corporations create internal capital markets, which are less prone to asymmetric information problems and, hence, they can sustain higher levels of debt. From this, we expect the sensitivity of long-term debt to cash flow to be higher for Wholly Foreign firms than Purely Domestic firms. As for the sensitivities of both short-term debt and long-term debt to collateral, we hypothesise that the sensitivities of both short-term and long-term debts to collateral should be lower for Wholly Foreign firms than Purely Domestic firms. One main reason accounts for this. First, because foreign firms are mainly 'alien' to the host economy, there exist some difficulties in the gathering and processing of information, making monitoring of foreign firms more expensive than monitoring of domestic firms (Doukas and Pantzalis, 2003). This means that WF firms that have little or no association with the host economy and are more susceptible to information asymmetries will be required to produce a higher value of collateral than local firms. Given that foreign firms' assets consist mainly of intangibles compared to tangibles (see Ethier, 1996 and Ethier and Hom, 1990), when the collateral of Wholly Foreign firms increase for instance, both short-term debt and long-term debt will increase but not as much as that of Purely Domestic firms. Next in our analyses, we consider the financing decisions of exporting and non-exporting firms. Apart from the contribution of Foreign Direct Investment (FDI), exports also contribute greatly to the Chinese economy. As indicated above in section 1.1, exporting firms are presumed to have sufficient financing to cater for their short-term projects because of the processes they have to undergo before they are able to export. They may however require more financing for their long-term projects. This therefore means that, like Wholly- Foreign Owned firms, exporting firms' short-term debt should be less sensitive to cash flow and their long-term debt should be more sensitive to their cash flow. The opposite is true for non-exporting firms. Again as explained in section 1.1, exporting firms because of the nature of their operations are endowed with more intangible assets. Therefore, using the same explanation provided for Wholly-Foreign firms, we expect both their sensitivities of long- term debt and short-term debt to be lower for exporting firms than non- exporting firms. We carry out similar analysis as we do for firms with different 22 degrees of foreign ownership. Specifically, we examine the sensitivities of both long-term and short-term debt to cash flow and collateral respectively, bearing in mind the other determinants of long-term and short-term debts for exporting and non-exporting firms. This is an innovation in this field of study. To the best of our knowledge, no study on capital structure, whether on China or in the rest of the world has ever looked at exporting and non-exporting firms. This is one of the things that make this study very unique. By doing this, we do not only get fresh results in this area of study but we are also able to compare these latter results with that of different degrees of foreign ownership. This will enable us draw conclusions about the financial behavior of globally- engaged firms. Despite the importance of firms in the growth and development of developing economies, until recently, work on capital market imperfections and foreign ownership have been based on anecdotal and survey evidence. Very little empirical work has been done in developing countries". One main reason for this is the difficulty in finding very detailed firm-level data for these countries. Even where some evidence exists, there is still some difficulty in explaining because these surveys are mainly done by institutions that give out loans such as the World Bank (Harrison and McMillan, 2003). This actually pushes some firms to say that they are financially constrained when they are actually not financially constrained. Our work becomes possible because relatively new and comprehensive panel surveys are now available for the chosen countries. 1.3Dataset Therefore, in order to investigate the relationship between financial constraints, firm ownership and crowding out in Ghana, we use data from a survey conducted by a team from the Centre for the Study of African Economies (CSAE) at Oxford University in collaboration with the Ghana Statistical Office (GSO). For Ghana these areas span the four largest urban centers in the country: This dataset contains comprehensive panel data set on a sample of "With the exception of Harrison and McMillan (2003), Harrison, Love and McMillan (2004) and Poncet and Hericourt (2008). 23 firms within the Ghanaian manufacturing sector. It was collected in five rounds over the period 1992 to 1998. The first three rounds were annual surveys, referred to as Waves I - III, as part of the Regional Program on Enterprise Development (RPED) organised by the World Bank, The fourth and fifth rounds of the data each cover two year periods. Firms that exited during the survey were replaced by other firms in the same sector that also had similar characteristics. The data was collected by a team from the Centre for the Study of African Economies (CSAE), University of Oxford, the University of Ghana, Legon and the Ghana Statistical Office. These surveys with a panel structure as well as the financial situation of firms will help provide more insight into the financing constraints for enterprises. One main limitation of this dataset is that it lacks a lot of financial variables and it is believed that this is mainly due to the reasons why the dataset was collected: It was mainly collected for to analyse issues about labour. For the work on China, we use a panel dataset from the ORIANA database. This database contains unique information on the financial position for public and private companies in the Asia-Pacific region. It is relatively new, comprehensive and underused. It is exclusive to Bureau van Dijk Electronic Publishing (BvEP) and combines data from various sources. BvDEP contains (where available) 24 balance sheet items, 25 profit and loss items and 26 standard ratios. In this dataset, all state-owned enterprises and other types of enterprises with annual sales of more than 650,000 US dollars are found in it. These firms operate in the manufacturing and mining sectors and cover all the 31 provinces in China. Also containing 110, 770 observations and over 20,000 firms, we believe it is the largest and most representative dataset in existence. It spans through the period 2000-2007. 1.4 Summary of Main Findings We find some very interesting results for our study using both datasets. For the work on financial constraints and crowding out on Ghana, we find that domestic firms are more financially constrained than foreign firms, but then foreign firms' borrowing has no impact on domestic firms' financial constraints. Although the former result is consistent with Harrison and 24 McMillan (2003), the latter is not. Also for the evidence on financial constraints on Chinese firms, while we find that purely domestic firms are more financially constrained than purely foreign firms, we find that joint ventures are not any different from purely domestic firms as far as financial constraints are concerned. As for the test of crowding out, we find results contradictory to what Harrison and McMillan (2003) find. Whilst Harrison and McMillan (2003) find that private domestic firms' financial constraints are worsened by foreign firms' presence, we find that private domestic firms' financial constraints are rather relieved by foreign firms' presence. This emphasises the importance of institutional differences and shows that certain situations are only peculiar to certain countries. With regards to our work on capital structure, whilst we find the sensitivity of short-term debt to cash flow and collateral respectively to be lower for purely foreign firms than purely domestic firms, we find the sensitivity of long-term debt to cash flow to be higher for purely foreign firms than purely domestic firms, For joint ventures, we find no difference between their sensitivities of both short-term debt and long-term debt to cash flow and collateral and that of purely domestic firms. The results for exporting and non-exporting firms are similar to that of the results for firms with different degrees of foreign ownership. Specifically, while we fmd the sensitivity of short-term debt to collateral to be lower for exporting firms than non-exporting firms, we find the sensitivity of long-term debt to just cash flow to be higher for exporting firms than non-exporting firms. This means that globally-engaged firms behave in a similar way as far as the sensitivities of the different leverage ratio measures to cash flow are concerned. As for the regional analyses, we find that the sensitivities of both short-term and long-term debt to cash flow and collateral respectively are higher for firms in the coastal region than firms in the non-coastal regions. 1.S Structure of Thesis The rest of the thesis is structured as follows. In Chapter 2, we include a literature review on finn financial constraints. This literature review is intended to serve both Chapter 3 and the Chapter 4. In Chapter 3, we provide a study on the financial constraints of domestic and foreign firms as well as a test on the 25 crowding out effect of firms in Ghana. Also as part of this study, we provide a background study on Ghana's financial system and the trend of FDI. In Chapter 4, we examine the same issues as in Chapter 3 but focus on the Chinese economy. Unlike Chapter 3, our data on the Chinese economy gives us the flexibility in investigating this issue on firms with different degrees of foreign ownership. Therefore, we are able to include joint ventures in our study. Also in this chapter, we brief ourselves on the Chinese economy by highlighting some important aspects of its financial system and FDI inflows. Given the high quality of our data and its disaggregated nature, we are also able to investigate this same issue in the regional context. Chapters 5 and 6 take a slightly different path but still focus on the Chinese economy. We turn to the issue of capital structure and foreign ownership. Therefore, in Chapter 5, we include its separate literature review. Regarding the main contents, we specifically investigate the leverage behaviour of firms with different degree of foreign ownership. Apart from this, we also investigate how the relationship between capital structure and its financial determinants vary across different ownership groups, exporting status and location. All these are done in Chapter 6. Chapter 7 concludes the thesis with a summary of our findings, some synthesis of the work on financial constraints of firms in the Ghanaian and Chinese economies, limitations of our study, policy implications which can be drawn from our study and some possible extensions to the study. 26 Chapter 2 2. Literature Review on Local Firms, Foreign Firms and Financial Constraints'" 2.1 Introduction The impact of financial constraints on investment decisions has been of much interest to economic researchers over the past decade. Empirical work on the relationship between financial constraints and investment has given controversial results, which have consequently led to a never- ending debate on the relationship. Fundamental to this line of thought is that capital market imperfections in credit and equity markets create a wedge between external and internal financing. If financial constraints do not exist, changes in internal financing should not have any influence on investment. In that respect, internal financing can be easily substituted with external financing which is more costly in the presence of informational asymmetries. 2.2 Explaining the Presence of Financial Constraints Using Investment-Cash Flow Sensitivity Fazzari, Hubbard and Peterson (FHP, 1988) were the first to provide a seminal paper on the monotonic relationship between investment-cash flow sensitivities and financial constraints. They split U.S manufacturing firms into three groups based on an a priori measure of financial constraints (dividend payout ratio), and compare the three different categories' responsiveness of investment to cash flow. They find that the firms with low dividend payout policy (that is, firms which are a priori likely to be financially constrained), have their investment exhibiting higher sensitivities to cash flow than firms that payout higher dividends, meaning that low dividend payout firms are actually more financially constrained. Kaplan and Zingales (KZ, 1997) criticize the work of FHP( 1988 ) by reclassifying the 49 low dividend firms (Class 1firms) that FHP(1988) classify 10 This literature review serves Chapters 3 and 4. 27 as financially constrained into 5 groups. According to them, although firms facing financial constraints should respond more to internal financing than unconstrained firms, it is not automatically true that a monotonic relationship would exist between financial constraints and investment-cash flow sensitivities. They classify the 49 low dividend payout firms based on the qualitative information in annual reports, quantitative information in the companies' financial statements and notes that describe the firms' financial status with where they plan to source funds from for future investment. Based on this information, the firms' degrees of financial constraints are determined. Their results show that the firms that are a priori likely to be financially constrained have their investment being less sensitive to cash flow than those that are a priori less likely to be financially constrained; these results contradict FHP(1988)'s study. In a response to KZ (1997), FHP (1998) argue that the classification scheme employed by KZ (1997) is not credible: it is complicated and relies on managers statements which may not necessarily be true. FHP (1998) also argue that KZ (1996)' s may reflect econometric problems II. 2.3 Investigating the Existence of Financial Constraints Using Ownership Structure A vast literature emerged afterwards with papers supporting either one of the above studies. However, most papers that evolved criticised the use of the dividend pay-out ratio as the basis for classifying firms into more or less financially constrained. Following this various studies emerged using different forms of classifications for firms. An almost unexplored dimension of this research area is the impact of foreign ownership on the firm's financial constraints, and also, the effects that foreign firms' borrowing from the local markets may have on domestic firms' financial constraints. Despite the importance of this issue, only few empirical studies have addressed it. Harrison and McMillan (2003), for instance, examine the differential effect of financial constraints on purely domestic firms and firms with foreign II 'Q' = the ratio of capital to the replacement of capital but KZ( 1997) use the ratio of market value of assets to book value of assets. 28 ownership. Although classifying firms into size, age, dividend payout ratios etc. have been the focus of most studies in this field, Harrison and McMillan (2003) introduce a new criterion for finn classification: domestic and foreign ownership. By employing data from Ivory Coast for 399 firms over the period 1974-1987 and augmenting the standard Euler investment model with financial variables, they specifically test whether both domestic and foreign firms face the same level of credit constraints and whether foreign borrowing crowds out domestic firms out of the local financial market. Following Whited (1992) the financial variables they use are coverage ratio, measured by interest/interest +profit and the ratio of debt to asset. These measures indicate the indebtedness of firms, and the firms' lack of collateral respectively. A high coverage ratio (or debt/asset ratio) implies the firm is highly indebted and therefore financially constrained. They also control for the lag of the dependent variable and its square and lagged sales over fixed assets. Their results indicate that financial constraints are more binding for domestic firms than for foreign firms. They then interact the financial variables with a measure of foreign firms' borrowing on the local capital market and interact the same financial variables with foreign sales. They nest both interaction terms into the same investment equation to properly investigate whether (I) foreign firms' borrowing on the local market exacerbates local firms' financial constraints, and (2) if they do, what may be the driving force. Their results indicate a positive and significant relationship between investment and the interaction term that contains the financial variables and foreign borrowing. Meanwhile, the interaction term with foreign sales and the financial variables remain insignificant, implying that foreign borrowing increases the sensitivity of the financial variables to investment. Harrison, Love and McMillan (2004) also examine the issue of whether capital transaction restrictions, in particular, Direct Foreign (FDI) impact negatively or positively on firms' financial constraints. By using a panel of 40 countries covering about 7000 firms from 1988-1998, they also use the Euler investment equation augmented by a financial variable. Rather than using coverage ratio and debt/asset, they choose to use cash stock to proxy for financial constraints. They interact the FDI variable with cash stock and control for a number of factors like the lag of the dependent variable, the ratio of sales 29 to capital and fixed effects. Their results show that FDI reduces firms' financing constraints, a result which contradicts the earlier findings of Harrison and McMillan (2003) which show that foreign ownership worsens the financial constraints of domestic firms. In order to introduce heterogeneity into their sample, they split the sample into G7 and non G7 countries. Their results show that FDI eases the financing constraints of the non-G7 countries, while the G7 countries are not financially constrained at all. Splitting the sample into firms with foreign assets and those without foreign assets suggests that firms without foreign assets face more financial constraints than multinationals, as the coefficient on FDI interacted with cash is only significant and negative for firms without foreign affiliation. In a related paper, Rutkowski (2005), first investigates whether domestically-owned firms in Central and Eastern Europe face higher credit constraints than foreign-owned enterprises, and second, if domestic enterprises' financial constraints are caused by incoming FDI. He undertakes this study using two alternative firm-level regression methods. The first, which is more direct, takes the ease of accessing the loan as the dependent variable (one for short-term loans and one for long-term loans), while the second approach, which is more indirect, tests if the relative level of investment is explained by some other financial indicators. Using data from the second edition of the 'Business Environment and the Enterprise Performance Survey (BEEPS II), this study provides evidence that domestic firms face higher credit constraints when applying for short-term loan than foreign-owned firms. The study also finds that FOI reduces foreign firms' financial constraints without increasing the constraint faced by domestic enterprises. In a panel study of manufacturing firms in Estonia over the period 1985-1995, Mickiewicz, Bishop and Varblane (2004) examine the link between financial constraints and investment. They test the hypothesis that the impact of financial constraints on firm investment does not differ across firms with different ownership status and size. One unique thing about this study is how financial constraints are interpreted. Rather than looking at the statistical significance of the financial variable in an investment equation, their main way of interpreting financial constraint is through the signs of the proxy for the degree of financial constraint; a positive sign on the proxy for financial 30 constraint implies that the firm is financially constrained and vice versa. Although they use the investment Euler equation as used by previous studies, they do not introduce any borrowing constraint into the Euler equation. They first estimate a basic investment model using the Euler equation with cash flow (measured by net sales/ previous year's capital), some dummies and previous year's investment/capital and its square (i.e. inv capital, = constant + inv_capitalt_1+ inv_capit_sqt_1 + nersalcapc, + year_dummies + sectoral contrasts + Eit ). Their results indicate a negative relationship between cash flow and investment implying that financial constraints are absent. This result is however not robust to the introduction of heterogeneity across different categories of firms into the equation in terms of ownership structure and firm size as the earlier results are altered. By splitting the sample into foreign and domestic firms and small and large firms respectively and interacting the cash flow variable with them, the results indicate a positive relationship between domestic firms and cash flow, and between the reciprocal of employment (small firms) and cash flow (small firms), which implies that a smaller percentage of foreign ownership (or firms of small size) makes a firm prey to financial constraints. According to their results, the absence of financial constraint initially found was based on the assumption of firm homogeneity in their estimation. They also find great evidence for financial constraints of firms which possess both characteristics of low foreign participation and small size. 2.4 Classifying Firms into whether or Not they are Globally- Engaged to Test for Financial Constraints Guariglia and Mateut (2010) look at the same issue of financial constraint and investment from a different perspective; digress from the use of fixed investment models augmented by financial variables by earlier studies and use inventory investment regressions to find evidence of financial constraints. Using a total of9381 U.K. firms, their sample was taken fromI993-2003. Their contribution to literature in this field is in two folds. a)To investigate whether firms which are globally engaged (measured by exporting firms and firms with a high share of foreign ownership) exhibit lower/no sensitivity of inventory investment to financial constraint indicators compared to their counterpart 31 firms which do not export or are domestically owned and b) classify finns using two different criteria based on traditional finn characteristics which divide firms into healthy and financially constrained firms and on global engagement status. Finns which fall below 75% of the distribution of sales, risk and age are classified as financially constrained whiles those with their distribution above are healthy. For global engagement the classification is based on firms' ability to export and ownership status. By modifying Kashyap et ai's (1993) Mix Variable (bank loanslbank loans +commercial paper), the financial variables they employ in the inventory equation is the ratio of short term debt to the sum of short term debt and trade credit (MIX I) and the ratio of the firm's short term debt to its current liabilities (MIX2), which is broader and takes into account other alternative sources of finance such as loans from insurance and financial companies, corporate bonds and commercial paper. This also differentiates this study from other studies in this field which have mainly relied on coverage ratio, cash flow and cash stock as proxies for financial variables. A positive and significant level of the mix variable implies financial constraint while a negative and insignificant mix variable implies firms are less financially constrained. Using an error correction model, Guariglia and Mateut (2010) find that inventory investment of firms which are financially healthy, based on the traditional criteria, exhibit lower sensitivity of inventory investment to the mix variables whiles financially constrained firms exhibit higher sensitivities. In particular, they find that the coefficients of the mix variables of the financially healthy firms are insignificant statistically and negative whiles those of the constrained group are significant and positive. Inventory investments of firms with global engagement (that is, those who export or have some foreign ownership) are less responsive to the mix variable than the inventory investment of those without global status implying that these firms are more financially constrained. In order to investigate whether global engagement mitigates financial constraints, constrained firms are split into those with global engagement and those without. Results reveal statistically significant coefficients for the mix variables for firms without global involvement but lower sensitivity of inventory to the mix variable for firms with global engagement. Consequently, participating in global activities attenuates 32 financial constraints. Their work therefore supports McMillan and Harrison (2003). Also, their results show that inventory investment is more responsive to firms which are relatively small, younger and more risky. In Guariglia and Bridges (2008)'s study, they examine whether global engagement may protect firms from financial constraint based on the failure probability of a firm, using a panel of 9420 newly established U.K firms from 1997-2002 with a probit model. They diversify the model by splitting firms based on their global status. Like Guariglia and Mateut (20 10), this classification of global engagement is based on two dimensions, exports and ownership so that firms who export and those with foreigners having a greater share are considered globally engaged and the opposite is true for domestic firms. They augment the probit model for survival probabilities with financial variables. The financial variables they use in this study are collateral; measured as the ratio of tangible assets to total assets, and leverage; measured by short term debt/total assets. The probability of a firm to fail is measured by a company whose position in a given year is in receivership, liquidation or dissolved. According to them, if global engagement protects firms from financial constraints, then the probability of firms to fail should be less responsive to the financial variables for globally engaged firms. On the other hand, firms without global status would have their probability of failure responding sharply to financial variables. In order to capture the effect that financial variables have on the survival probability of a firm, they interact the financial variables with globally engaged firms and purely domestic firms. In their first estimation, they consider globally engaged firms (that is, firms that export and/or are foreign owned) jointly without splitting the firms into those that export and those that are foreign owned. Their results show that financial variables'( leverage and collateral) have greater effects on firm's probability to fail for purely domestic firms than firms with some global status meaning that global engagement alleviates firms from financial constraint and consequently 'death '.The signs of the coefficients of the financial variables also support theory. A higher collateral and lower leverage are associated with low failure probabilities of firms. Using the Cox proportional hazard model and the random effect logit model does not change the results. In order to investigate whether the initial 33 findings would hold for the individual dimensions of global engagement they consider in their studies, they estimate another logit equation with each of the two dimensions in the same equation interacted with each of the financial variables but the results does not change. In addition, introducing old firms into the sample (that is firms before 1996) does not change their results. 2.5 Examining the Link between Investment and Financial Constraints In Periods of Currency Crises Desai, Foley and Forbes (2004) investigate the link between financial constraints and investment in a unique way, by looking at how firms with multinational affiliates and local firms respond differently to large depreciation of a currency. Using a sample of 25 emerging markets from 1991-1999, their results indicate that in the years of and after depreciation, firms with foreign affiliates increase sales, assets and investment more than local firms. In their view, currency crises can either make firms very competitive on the international market by promoting exports, or increase firm's indebtedness due to large borrowing from foreign sources without corresponding adjustment of capital structure. Hence, the effect of financial constraint on firm investment in periods characterising depreciation can be seen through these two ways. In order to examine why differences exist in the response of multinationals and local firms to currency crises, they examine the impact of depreciation on relative changes in operating profitability which measures competitive benefits, and leverage (current liabilities/ long term assets), measuring financial constraints. Their results indicate that the operating profitability of both domestic and foreign-owned firms respond in the same way to currency crises, but the leverage of local firms responds more to currency crises than multinationals', suggesting that local firms' inability to increase investment in times of currency crises is due to financial constraints. In a similar study, Blalock et.al (2008) examine whether capital imperfection restrains firms from investing during currency crises. Using data from the Republic of Indonesia's Budan Pusal Statistik, they find that the value-added and employment for exporters increased after the 1997 Asian 34 crisis, implying that, Indonesian exporters took advantage of the devaluation process. However, their results show that only exporters with foreign ownership increase their capital significantly. This suggests that financial constraints prevent domestic firms from benefiting from the terms of trade. In our study, we investigate the financial constraints of foreign and domestic firms and the test whether domestic firms in Ghana and China are crowded out by foreign owned firms. Different from the aforementioned studies, our analysis on Ghana investigates the financial constraints of foreign and domestic firms using a totally different category of foreign firms. These foreign firms may not necessarily be FDIs as seen in previous studies. Most of these firms are owned by foreigners who have lived in the country for so many years and have established themselves within the country. They are mostly from Asian origin. The firms under investigation are mostly firms that have been passed on from one generation to another. Owners of these firms have formed networks with their colleagues and can also receive some form of assistance from home. As a result, we believe that results from this study will be able to confirm whether being foreign-owned is what makes the difference. As for or analysis on China, we investigate this same issue using firms with different degrees of foreign ownership. All the studies reviewed above divide firms into either domestic or foreign. This does not allow a complete capturing of the financial constraints of joint ventures. We fill in this gap by specifically dividing firms into three; purely domestic, joint ventures and wholly-foreign owned firms. By this we are able to separate joint ventures from purely domestic firms and wholly-foreign firms to capture the financial constraints of firms with different degrees of foreign ownership. In addition to this, we carry out regional analyses to find out whether the level of financial constraints of firms differs across the different regions of China. No study has done this as far as we are aware. We start our analyses in the next chapter on Ghana. 35 Chapter 3 3. Foreign- owned Firms and Financial Constraints: Evidence from the Ghanaian Manufacturing Sector 3.1 Introductlen'? The Ghanaian economy has in recent times been doing pretty well with unprecedented growth, both from commodity booms and improved macroeconomic management. It managed growth of over 6% and inflation of just over 10% for a number of years. From an average growth rate of about 4.4% in the early 1990s, its average growth rate increased by just about 0.1% in the late 1990s. The economy started picking up in the early 2000s when its average growth rate surged to about 5.6%. Interestingly, in 2008 for instance, despite the high food prices and the soaring crude prices that hit the economy, it still was able to maintain some growth resilience as the officially recorded growth rate was about 7.3%13.This growth rate was even higher than the rate for 2007. This increase in the GOP growth rate translated into a higher real per capita GOP growth rate of 4.8% up from 4.1% in the same year. Considering the sectoral growth rates, the industrial sector, comes second after the service sector. Whilst the growth rate for instance for the industrial sector was about 8.1% in 2008 (a rise of about 0.7% in 2007), the growth rate for the service sector was about 9.3% (a rise of about 1.1%). In terms of industrial output, the manufacturing sub-sector has mainly been its major contributor. From1990 to 1999, it accounted for about 58.2% of industrial output. For its share in GOP, it appears to be on the same line with the construction sub-sector, both contributing large shares to GOP as far as the industrial sector is concerned. Further, out of the GOP growth rates that have been stated above, the manufacturing sector has been one of the major contributors as far as the 12 All figures in this section unless indicated otherwise come from the State of the Ghanaian Economy, various editions I3See figure 3.AI in appendix for details 36 contribution of the industry sub-sector is concerned. It has been movmg alongside with the construction sub-sector. In the early 1990s, its share of real GDP was on average about 8.6%, compared to the mining and the construction sub-sectors that contributed on average 1.4% and 2.9% respectively. It then increased its contribution to about 9.1% in the late 1990s. However, its contribution fell to about 8.6% in the early 2000s and further to about 7.9%, lagging behind the construction sub-sector in 200814. Despite the manufacturing sector's valuable contribution to the economy, figures indicate that it has been experiencing the least growth rate compared to the other sub-sectors. Its growth rate hovered around 2.2% on average in the 1990s compared to an average growth rate of about 6% in the construction sector. Between 2000 and 2008, though its growth rate increases to an average of about 5.9%, it still couldn't match the construction sector. The construction sector had an average growth rate of about 7.3%.15 An interesting question to pose is why the sector is not growing. Can it be attributed to financial constraints then? From the State of the Ghanaian Economy (various years), the amount of credit to the manufacturing sector (at 2006 Ghana cedis) decreased from 647.7m Ghana cedis in 1993 to about 375.25m Ghana cedis in 1997. In 2000, it increased to about 1528.359m Ghana cedis. It however declined again to 890.94m Ghana cedis in 2004 and further to about 619.68m Ghana cedis in 2008. This situation is likely to cause firms to be constrained financially. The case might even be worse for domestic firms who are basically limited to the domestic credit market. Foreign firms located within the country can at least have some credit from the international market, other foreign companies or even their parent companies in the case of multinationals if the domestic credit market fails them. Even if credit is made available on the domestic market, foreign firms are likely to have better access to these credits unless some specific directives exist in the economy for domestic firms. In the case of Ghana, however, this is not so. Banks lend only to firms using financial consideration and not by political consideration. Following this argument, many developing nations often set out policies to attract foreign investors to help relieve their domestic firms of their 14Seefigure 3.A2 in appendix for details ISSee figure 3.A3 in appendix for more details. 37 financial constraints. But does this effectively occur? Essentially, if these foreign investors come in with scarce capital and generate technological spillovers, they may provide support to domestic firms. This may in tum promote economic growth and development in the host country. With these benefits in mind, Ghana is now beginning to attract significant levels of FDI as new policies are aiming at making foreign investment an integral part of the macro economy. For instance, net foreign direct inflows increased from about 14 million dollars in 1990 to about 106.5 million dollars in 1995 and further to about 144.97 million dollars in 2000. By 2007, it had increased to about 970 million dollars (WDI, various yearsj'". As to whether just attracting foreign investment into the country would generate a non- trivial impact on the whole economy is one argument on its own. When foreign investors come in to borrow from the domestic financial institutions, then, because they typically have good reputation and are considered more productive and less risky, they become a better bet for financial institutions. As a result, they are able to compete effectively with domestic firms and deprive them of the limited capital available (crowding out effect). Limiting the situation to Africa, most African developing countries have not been able to enjoy the benefits of having a stable investment inflow. Expectations that Foreign Investment could be a channel through which economic growth and development could take place have not been fully achieved; Ghana could be no exception. The latter argument could explain the lack of impact ofFDI on growth. Bearing these arguments in mind therefore, the overall impact of foreign investment on domestic firms, and for that matter, the whole economy depends on which of the two outweighs the other. Perhaps, the insignificant impact of foreign investment on Ghana's economy can be attributed to the fact that the crowding out effect dominates the benefits of attracting foreign investment Using data from the Centre of the Study of African Economies (CSAE), we examine whether, first, domestic firms in Ghana face more financial constraints than foreign firms, and, second, whether foreign firms' borrowing from the local financial market worsens the financial constraints of domestic 16 WDI represents World Development Indicators. 38 firms. It is worth-noting that this dataset which covers the period 1991-1997 marks the period after financial development in Ghana. By undertaking this study, we may be able to examine the impact of financial development on the economy as we expect more investible funds to be available to firms due to the increase in the number of banks. Although this dataset is the only most comprehensive finn level dataset on Ghana, it lacks a lot of financial variables: Thus, making it difficult to carry out a lot of financial analyses. We combine ideas from Whited (1992) Bond and Meghir (1994) Gilchrist and Himmelberg (1998), Love (2000) and McMillan et al (2003), to construct a model that relates adjustment costs in different periods. We test the predictions of this model using, an Euler Equation augmented with the financial variable, debt. As in other investment literatures, we interpret the coefficient of debt as the degree of financial constraints faced by firms. Basically, a significant coefficient of debt will mean the finn is fmancially constrained. Conversely, an insignificant debt variable will imply that, when there is a change in debt, the finn can substitute it with another form of finance, making it financially healthy. In the context of the Euler equation, if firms are not financially constrained, then changes in debt should not affect future investment. We show how the coefficient varies across different groups of firms classified into foreign-owned and purely domestic. We also explore whether borrowing by foreign firms worsens the financial constraints of domestic firms by introducing the variable, share of foreign borrowing which we interact with debt. As far as we are aware, there are only few empirical studies that have tackled this issue. Alexander Rutkowski (2005) for instance, investigated the impact of foreign firms' borrowing from the local capital markets on domestic firms' financial constraints by focusing on Central and Eastern European Countries (CEEC). Harrison and McMillan (2003), also, investigated this issue by concentrating on Ivory Coast, but, may not provide a full picture about what happens in Africa. We therefore examine this same issue in another African country: Ghana. By doing this, we will contribute to existing literature in two ways. First, we will be able to make generalisations about the impact of foreign firms' borrowing on the financial constraints of domestic firms in West Africa. Second, we will also be able to make a direct comparison between these two economies as similar techniques, variables and methodologies will be applied. 39 Our choice of Ghana for this study is motivated by two main reasons. First, as far as we are aware, virtually no study has examined this issue in Ghana. We consider this study somewhat important in the context of Ghana. This is because in recent times, the government has been making policies to attract foreign investment as it believes foreign investment could be a channel through which growth could be achieved. Therefore, the outcome of this study will help the government determine whether these policies are worth implementing. Second, our dataset offers us the opportunity to examine this issue using another form of foreign ownership which is not the same as the traditional FDI, but which we believe could be interesting to investigate. One interesting and clear distinction can be made between our study and that of Harrison and McMillan (2003). In our dataset, foreign- owned is not directly equivalent to FDI as traditionally understood and described by Harrison and McMillan (2003). Foreign-owned in our dataset refers mostly to ownership by individual foreigners that are not originally from Ghana. These individuals often belong to the Lebanese and Asian communities who play a very significant role in business in Ghana. Many of them have become third and fourth generation Ghanaian citizens and own most of the big supermarkets in Ghana. Majority of the hotels, restaurants and fast food companies are owned and run as joint ventures between Ghanaians and Lebanese, and Ghanaians and some Asian communities, especially China and India. Essentially, because of their networking with their communities, these firms which may either be fully owned by foreigners or may be joint ventures may have access to various forms of financing, making them relatively less credit constrained than domestic firms. Yet, they may be more credit constrained than the traditional FDI since their source of external financing may be restricted to either their own communities or the domestic market as they do not have parent companies. In this chapter, we assume that ownership by individual foreigners serves as a representative for FDI and can capture important features of the differences in the financial constraints behaviour of foreign and local firms. If these firms who are not actually FDIs are, one, less financially constrained than the domestic firms, then, we can tell a good story about foreign ownership. 40 Using this group of foreign-owned firms and domestic firms in our analyses, we find results consistent with Harrison and McMillan (2003). Specifically, we find that domestic firms face a relatively higher degree of financial constraints than foreign firms. However, contrary to Harrison and McMillan (2003), we find that, domestic firms' financial constraints are not affected by foreign firms' borrowing. For policy purposes, we recommend that Ghana's recent attempt to attract foreign investment should be encouraged in a way, as these foreign firms do not negatively affect domestic firms. In attracting them, the economy could derive other benefits that foreign investors provide such as technological spillovers. The remainder of this chapter is organised as follows: Section 2 provides an overview of Ghana's Financial System and Foreign Direct Investment in Ghana. In section 3.3, we describe the data. In section 3.4, we present the various approaches to testing for the existence of financial constraints, our categorisation method and the variable measurement. We present the methodology and descriptive statistics in sections 3.5 and 3.6 respectively. We show our results in section 3.7 and provide robustness check. In section 3.8, we conclude and provide some recommendations for policy. 3.2 An Overview of Ghana's Financial System and Foreign Direct Investment 3.2.1 Ghana's Financial Sector 3.2.1. IPre- Financial Sector Reforms Era Prior to Ghana's independence, its colonial masters imposed its power on the financial sector. In 1875, the British government successfully monetized Ghana (then known as Gold Coast). During this period, they tried to restrict the use of currency in Ghana to just the British currency. For example, in 1880, a demonetization exercise was carried out to allow the lawful withdrawal of all other currencies in use in order to limit the currency to just gold and British sterling. This led to the establishment of the Government Savings Bank in 1888. The British Bank of West Africa with its headquarters in Nigeria 41 extended its territories to Ghana in 1896. In 1917, the Barclays Bank (formerly known as the Colonial Bank) was established to encourage competition in the banking industry. Problems which the native people faced in accessing credit as a result of the so called "discriminatory practices" exercised by the foreign commercial banks as well as the high collateral demands, together with the fact that certain sectors in the economy regarded as important needed specific financial institutions to take care of their needs motivated the government to set up its own commercial and development banks. The Bank of the Gold Coast (now Ghana Commercial Bank) was then set up in May 1953 as a native bank to further boost the activities of the two commercial banks already in operation. Massive pressure mounted on the colonial government by the native people led to the establishment of the Bank of Ghana on March 1 1957, a few days before Ghana gained independence. To Ghana, the establishment of the Central Bank meant that it had gained financial independence. Upon the request of the local people the Bank of Ghana was given the total mandate to implement policies which were deemed fit to promote growth in the economy. Later, three commercial banks, The National Investment Bank, Agricultural Development Bank and the Bank for Housing and Construction were set up to support the activities of the Ghana Commercial Bank. In 1975, the National Savings and Credit Bank, known previously as the Post Office Savings Bank and the Cooperative Bank were established to provide credit to small-scale borrowers. Ghana's desire to develop rapidly after independence led to the implementation of certain policies in the economy that resulted in financial repression later on. The government owned all the banks which were established and even possessed about 40% equity in the two expatriate banks which were already in existence. The Bank of Ghana controlled extensively interest rates, and credits were allocated to specific sectors in the economy. These policies were designed to achieve some desired level of investment, ensure proper allocation of credits among various sectors of the economy, and to keep interest rates generally low. Unfortunately, the policies implemented could not achieve the targets set. Rather, the economy faced financial repression with interest rate hitting negative levels. 42 Worse of all, to reduce the amount of money in circulation, in 1979, the government imposed a 30% tax on all cash holdings outside the banking system below five thousand cedis and a 50% tax on the cash holdings above five thousand cedis. In addition, in the early 1980s, the Central Bank, withdrew all fifty cedi notes from the system and replaced them with smaller denominations. Those from whom the fifty cedi notes were withdrawn did not receive any compensation immediately. When compensation was due, most of them had misplaced their receipts. In short, this period witnessed severe financial repression and people lost confidence in the banking system. Most of the policies implemented could not achieve the targets set. The supply of credits including those to the priority sector was reduced and aU the banks but the ones with foreign shares went bankrupt. The severe financial repression together with the loss of confidence that had subjugated the financial system in the Ghanaian economy led to the implementation of reforms in the latter parts of the 1980's. It was not until the initiation of the Financial Sector Adjustment Program (FINSAP) that it was realized that the financial sector was extremely distressed. 3.2.1.2 Financial Sector Reforms The financial reforms which took place 10 Ghana constituted institutional restructuring, improvement in the legal and regulatory framework for banking operations, and interest rate liberalization. 3.2.1.2a Public Sector Banks Restructuring The reforms of the financial system started with restructuring of the public sector banks in 1989. Their focus was on the balance sheet as most banks before the reform took place had gone bankrupt. 3.2.1.2b Financial and Managerial Restructuring Major restructuring took place during the reforms. The organizational structure that took place consisted of both management and financial restructuring of the banks. With regards to management, there was reshuffling in most of the banks especially at the management level. Financial restructuring according to Sowa 43 (2003) involved mainly recapitalization of the banks with equity injection where liquidity was low. Banks' balance sheets of non-performing assets were also cleaned up. 3.2.1.2c Institutional Restructuring The structure of the financial system was changed. New institutions were set up to shake-up the financial system. Some banks were merged and some sold out. Actually, five banks and twenty non- bank institutions were set up under FINSAP. For example, in 1991, the Security Discount Company (SOC) was established mainly to eradicate the monopolistic power of the Consolidated Discount House which was established in 1987. Many more rural banks were created. It was speculated that the commercial banks had been unable to extend its activities to the rural area. The main aim of creating the rural banks was to fill the gaps created by the commercial banks. 3.2.1.2dCapital Market FINSAP brought to fore the Ghana Stock Exchange. The Ghana Stock Exchange listed 12 companies and one Government Bond when it started its operations in November 1990. The value of listed shares amounted to 43 billion cedis in 1992 as compared to 30 billion cedis in 1991. According to Sowa (2003) ever since the inception of the Ghana Stock Exchange, the market has mobilized about 140 billion cedis and US 4.8 million dollars through equities and bonds and the number of listed companies has risen to 21. 3.2.1.2eInterest Rate Liberalisation Under the Structural Adjustment Program, market forces were allowed to determine interest rates freely. Government involvement in interest rate determination was cut off. The deregulation of interest rates began with the eradication of the maximum and minimum deposit policy in September, 1987. Subsequently, the commercial banks were given the mandate to set up their own rates in March, 1989 following the eradication of minimum lending rate in February, 1988. By 1990, the policy that allowed the agric sector to have 44 access to a compulsory lending of about 20% was also abolished. Contrary to the pre-reform era, the reform was able to move both Treasury bill rate and real lending rate from negative to positive. 3.2.1.3 Operations of Ghana's Financial Sector Ghana's financial sector is characterised by both formal and semi formal financial institutions. The formal financial institutions include banks and co- operative societies, and the semi-formal financial institutions include susu groups, susu collectors and money lenders. Currently, the banking sector in Ghana is made up of the Bank of Ghana, eight Commercial Banks, three Development Banks, three Merchant Banks and one hundred and thirty-three rural banks. Apart from the rural banks, the distribution of the other banks is skewed towards the urban areas, which are mainly found in the southern part of the country. For instance, out of the thirteen districts in the Northern Region, seven of them have no banks. This gives the available banks a large client base. The banks are characterised by relatively high value and longer duration loans, which call for a lot of bureaucratic processes and collateral. As a result, the majority of the people in the rural areas and the northern part of the country engage in semi-formal financial practices. Susu groups are mainly savings groups. A modified form of the susu group is the Rotating Savings and Credit Association (ROSCA). The ROSCA is made up of a group of individuals who come together and make regular contributions to a common fund, which is then given as a lump sum to one of the members in each cycle. This means that a member will lend money to other members of the group through his regular contributions to the fund. The total amount collected in each cycle is given to each member in tum. The person who gets the lump sum in each cycle is determined either by common consensus, lottery, or bid. By consensus, the members normally decide to give the total amount to the one who needs the money most. Where they ballot, members who have already received the lump sum contribute, but do not participate in the lottery till the cycle ends. With bidding, the highest bidder 45 takes the lump sum minus the bid. ROSCA's are flexible and can easily be used in both the rural and urban areas. The Susu collectors agree on how much savers would want to save with the savers themselves, and visit them at regular intervals for collection. They deposit the money at a bank, invest the money in their businesses or lend it to others for a period of time. At the end of the period, the susu collectors return the savers' deposit less a small commission. The Ghana Co-operative Susu Collectors Association oversees the activities of susu collectors. The moneylenders are usually the rich farmers or traders who lend out their own funds or have access to credits from the banks. They usually have good knowledge of their borrowers and so discourage fraud. Their focus is not mainly on how borrowers use the money they borrow but on their ability to repay the loans. As opposed to the formal financial services, the semi-formal financial services have low transaction cost, door-to-door services, and the ability to adjust to changing circumstances. In Ghana, people may also borrow from informal sources such as friends, relatives and religious groups. These often accrue no transaction costs and do not take any formal contractual form. By establishing a cross-border relationship with foreign firms, domestic firms can expand their credit options. This may in tum enhance the growth of domestic firms and the economy as a whole. Quite recently, Ghana's policies have been geared towards attracting foreign investment. 3.2.2 FDI Trends and the Macroeconomic Environment in Ghana Ghana, a country endowed with natural resources looked very promising and bright when it gained independence. Disappointingly, it has been unable to attract significant FDI flows. When Ghana gained independence in 1957, a socialist policy was adopted with the intention of narrowing the gap between the rich and the poor in the country. This initially put the economy at the fore front of Sub- Saharan African countries in terms of economic growth. Unfortunately, this policy was misconceived and its positive impacts remained unsustainable. By 1960, the economy had already started experiencing some downturns in its macroeconomic indicators. From 1966-1969, Ghana's 46 economy was revitalised when policy-makers at the time solicited the assistance of the IMF. Among the policies implemented were divestiture, devaluation and trade liberalisation. Macroeconomic mismanagement coupled with political instability which characterised the country in the 1970s and early 80s pushed the economy into severe crisis. In fact, for the short period between 1972 and 1981, four different military governments took over the country. Turning to the macroeconomic environment, the highest GOP growth rate was registered in 1973 (15.25%) and this trickled all the way to negative (- 12.85%) rate in 1975, recording the lowest GOP growth rate in that period (Aryeetey et aI., 2004).Annual inflation hovered around double-digit figures and, by February 1984, it had reached 126% (Bank of Ghana Statistical Bulletins, various issues). During this same period, the country was heavily indebted and financed its indebtedness through banks and seignorage leading at negative interest rates (Aryee