During the past 40 years one of the most controversial issues in finance theory has been the theory of capital structure. The modern theory of capital structure have been developed and expanded on the basis of seminal work by Modogliani & Miller (1958), which suggests that firm's market value is independent with its capital structure built on assumption of perfect capital market, and hence equity and debt can be substitutes for each other. Moreover, the key assumption under the M&M proposition is to assume exists perfect capital structure, which includes no asymmetrical information, no tax, no translation cost and so on. So far, most of researchers have been focus on test the validity of modern theory of capital structure as well as in different business sectors. For instance, Modigliani & Miller (1996) focus on analysis electric-utility companies and Titman & Wessels (1998) used manufacturing firms as a sample. However, even after long time intensive research, there is still no single theory have been found to solve the capital structure puzzle (Bhaduri, 2002). Furthermore, the empirical researchers have failed to determine the firm specific determinants of capital structure when the theories of capital structure are quite abstract and not directly observable (Titman and Wesswls, 1988).
Once the assumption is invalid in the real business world, the choice of capital can affect firm's value as important factor. The different combination of debt and equity will affect firm value immediately. Therefore the firm manager is very necessary to find best combination of debt and equity in order to maximum firm value. It becomes increasingly aware that the development of alternative theories of capital structure decision and their empirical analysis is extremely important. Now, researchers have already realized that the choice of capital structure between debt and equity which also depend on firm characteristics.
Since different countries have different economic systems, then formation of a different capital structure. In the long time, UK has been development under the traditional and laissez faire capitalist economic system. It has highly developed securities market and has a power to against the traditional gathering of financial factor. Moreover, such traditional economic system is formatted by a company's financing characteristics.
The UK can be characterized as having a broadly similar financial and legal environment to the US. It has a common law legal system with good investor protection, well-developed financial markets and an active market for corporate control. Bank finance and inter-company ownership relationships play relatively smaller roles than in some countries. The most obvious differences between the UK and US relate to tax and bankruptcy codes and the size of the corporate bond market (Rajan and Zingales, 1995).
There are three major reasons to choose UK as a sample country. So far, a number of researchers have documented the firm specific determinants of capital structure which based on firms in the US. Recently, researchers have shown an increased interest in to document international results. Moreover, the growing numbers of empirical evidences of Chinese firms have been added to support previous studies. However, only few studies have used UK data, which includes Bennett & Donnelly (1993), Ozkan (2001) and Bevan & Danbolt (2002, 2004). There is no more available evidence for UK firms after last publish paper in 2004. On the other hand, there are some conflicts in the empirical results. Therefore, this study attempts to reduce the time gap from 2007 to 2011 and add more evidence for UK firms. Secondly, the target firms listed in London Stock Exchange and represents almost 81% of UK market capitalization. Compared with other unlisted firms, these firms should have a relatively stable capital structure and have their own capital structure. Based on the previous studies, we found that uniqueness and liquidity is lack of empirical evidence in the context UK. Therefore, this study will include uniqueness and liquidity as explanatory variables to provide new evidence of firm specific determinants of capital structure. Finally, UK as developed countries which have a complete financial system, it will much easy to do research in this background. Furthermore, the existing theories of capital structure and theoretical models have more explanatory power and suitable in developed countries.
1.2 Research objectives
The objective of this research is to investigate whether there is a significant relationship between firm characteristics and capital structure. We selected seven firm characterizes Based on empirical evidence on capital structure, and use selected the target firms listed in FTSE 100 over the period from 2005 to 2011. The following research question will be help to achieve our objective.
1. Which firm characteristics can affect UK firm's capital structure?
2. How do firm characteristics affect capital structure?
1.3 Outline of Structure
This dissertation is organized into six chapters as follow. Chapter 2 reviews the relevant theories of capital structure and documents the empirical evidences based on the United States, United Kingdom and China data and brief discuss the firm specific determinants of capital structure in last section. Chapter 3 then proposes the research methodology in this study and develops testable hypothesis according to the theories and literature presents in chapter 2. Chapter 4 represents the result of analytical discussion. The final chapter 5 concludes that summary the main findings, limitation of the study and recommendations for future research.
1.4 Summary of Chapter
This chapter brief discusses the background of theories of capital structure. And it explained the necessary of extent the research in UK listed firms. Also, this chapter presents the three major reasons to choose UK as a sample country. Moreover, the capital structure of UK firms has been analyzed. This objective of this study is investigated whether there is a significant relationship between firm characteristics and capital structure. In order to achieve the objectives, there were two research questions have been added. Therefore, in the next chapter, we will reviews the relevant theories of capital structure and empirical evidences on capital structure which based on the firms in USA, UK and China.
Chapter 2: Literature Review: Theoretical background and Empirical research
2.1 Introduction
This chapter is divided into two sections. The first section is theories background that reviews the primary theories of capital structure. It includes M&M theory, trade-off theory, pecking order theory, agency theory and market timing theory. The second section is empirical research on firm determinates of capital structure in developing countries and developed countries.
2.2 Theories of Capital Structure
The early capital structure theory is actually preliminary. The start of the modern theory of capital structure could be regarded as the seminal work which was done by Modigliani and Miller in 1958. In this paper, they supported the net operating income approach and rejected the traditional theory of capital structure. Moreover, they proposed the first capital structure theory, which is referred to as M&M Proposition. This proposition I states that the market value of the firm is independent of its capital structure which based on the assumption of a perfect capital market (Modigliani & Miller, 1958). It assumes there is no asymmetrical information, no tax and transaction costs, nobody can individually influence market prices, there is s single interest rate for borrowing and lending, and there are homogeneous products (Modigliani & Miller, 1958). Moreover, M&M Proposition I suggests the value of geared firm is equal to ungeared firm, which means a firm's capital structure do not affect the firm's total value. (Brealey, Myers, & Allen, 2008). Furthermore, the MM Proposition II states that the capital structure does affect the expected rate of return on the common stock, which is built under perfect capital market assumption, expect no tax.
In this theory, the assumption of a perfect capital market is very important. The assumptions should not be relaxed, otherwise the result might be change: a firm's capital structure would become relevant to its overall value. Consequently, as MM theory has discussed the capital structure could be irrelevant from the firm's value under what conditions, so that others could find out the capital structure would be relevant to the firm's value under what conditions. Since the market in the real world is far from perfect, many researches have contributed to describing the relationship between a firm's capital structure and its value by criticizing MM theory's assumptions.
2.3 Trade-off Theory
Trade off theory is basically derived from the M&M theory, which initially developed by Modigliani and Miller (1963). They suggest debt can stimulate firms to presume tax shield. Theoretically, in order to gain maximization of tax advantages, firms intent to borrow as much of debts to strengthen their profitability. However, the higher level of debt the higher possibility that firms may financial distress concerns firms failed to meet its obligations. Modigliani and Miller (1963) argue one of the advantages of using maximizing level of debt concerns reducing interest payment from corporate tax. While Kim (1987) point out the potential cost of financial distress is closely related with high level of leverage. Warner (1997) state financial distress plays a crucial role in theories of capital structure, and increasing costs of financial distress have significant effects on firms operations and decision makings, while financial distress costs are commonly refers to bankruptcy costs. Megginson, et al (2007) argues bankruptcy costs can be categorized as direct cost and indirect costs. Direct costs related to the spending which directly associates with bankruptcy administration, such as lawyer fees, or fee paid to investment banker and accountant. On the other hand indirect costs are incurred from the consequences of bankruptcy, the costs like losing of important employees and sales during bankruptcy are all belonged to indirect costs. Therefore, on the basis of above arguments, bankruptcy costs can have great impact on firm's capital structure, and this has provide intensives for many firms to seek the optimal level of firms leverage, meanwhile it also lead to the key concept of trade off theory that is the tradeoff of between tax benefits and financial distress.
A number of empirical researches are conducted which challenge the trade- off theory. For examples, Myers (1993) discovers some evidences associates with trade off theory, he have found the correlation between financial leverage and firm's profitability are inversely related. According to Bradley et al (1984), the findings of their empirical research indicates financial distress costs are inversely correlated with firm's leverage, besides they also find out the non-debts tax savings are inversely associated with firm's leverage. To the contrary, there are several evidences which supports the trade-off theory. Byoun (2008) have found some evidences which accordance with the trade-off model, shows firms can adjust target leverage level under asymmetric method. Graham and Harvey (2001) investigate the principle of trade off theory through survey firm's capital structure, the finding of their research suggests majority of firms have their target debt ratios which consistent with trade off theory.
2.4 Pecking Order Theory
The pecking order theory is also an important theory on capital structure. Donaldson (1961) has first stated this theory. After that, it was developed by Myers and Majluf (1984) that a firm follows a financing hierarchy, which means the firm should prefer retained earnings as internal financing rather than external financing on the basis of the their studies. Moreover, if the firm must use external financing, they will issue the safest security first, then perhaps equity as a last resort when the company runs out of debt capacity (Myers, 1984).
The pecking order theory is usually on the basis of the idea of asymmetric information, which suggests that to compare with outsider investors, the mangers know more about their companies' future potential, risks and value (Brealey et al 2006). Furthermore, asymmetric information can be key decision in the choice between internal and external financing, and issue new debt or equity securities of firm (Brealey et al, 2003).
According to Baker & Martin (2011), the empirical evidence associate with the pecking order theory is mixed. Shyam-Sunder and Myers (1999) uses a simple empirical mode to analysis the sample of 157 US firms, they find that all of these firm act to fund their financing deficits with debt and the pecking order theory provides a very good guide for their financing behavior. Consistent with this view, According to Fama and French (2002), in small and high growth firm, the pecking order theory could perform better than others. On the other hand, some of authors still have found this theory is also regarded as a reasonable descriptive empirical mode of corporate capital structure. Bunn and Yong (2004) taking a similar approach to Frank & Goyal, but using UK date, found clear evidence to support the trade-off theory.
2.5 Agency Theory
Moles et al (2011) concluded that both the trade-off theory and the pecking order theory offer some insights into how managers choose the capital structures for their firm. However, other factors such as agency must be taken into account to consideration. Therefore, in order to achieve optimal capital structure of firms, Jensen & Meckling (1976) presented agency theory, which was a significant contribution to the theory of capital structure.
Agency cost is defined as 'the cost to the principal (shareholders) caused by agents (managers) pursuing their own interests instead of the principal's interest (Stickney et al, 2009). Because of the different interest between principal and agents and separation of control and ownership in a firm, all of these cause conflicts. Jensen & Meckling (1976) identify two types of conflicts. The first one is the conflicts between shareholders and managers. This conflict occurs due to the managers and shareholders have different personal interests pursued (Baker & Martin, 2011). In particular, managers might be more interested in maximizing their own wealth rather than increasing the wealth of shareholders and the firm value (Schroeder et al, 2010). Under this scenario, the managers may tend to taking negative NPV projects which might only expand the firm but damage the wealth of shareholders. Furthermore, there also would be a conflict between debtors and shareholders. (Jensen & Meckling, 1976). This conflict arises because the debt contract gives equity-holders an incentive to invest 'sub-optimally' (Harris & Raviv, 1991). It means that debt holders can transfer risk to equity holders by investing in risky projects (Jensen, 1986). For instance, if the project successful and with large return, the equity holders would take most of the gain. Nonetheless, if the project fails, equity-holders only holds limited liabilities, but debt-holders would have to bear the consequences (Harris & Raviv, 1991).
Harris & Raviv (1991) noted that agency costs are important determinant of firms' capital structure, because it analysis types of agency costs which can help explain the relevance of capital structure. The first type pf agency cost is the 'underinvestment or debt overhang problem' (Myers, 1977). Brigham & Ehrhardt (2010) defined underinvestment problem as ' a type pf agency problem in which high debt can cause managers to forgo positive NPV projects unless they are extremely safe'. According to Brealey and Myers (2000), almost all firms' capital structure could be affected by the underinvestment problem, and it is also significant in high debt ratio firms since financial distress. Another agency cost arises when managers select project only depend on their personal benefits. The mangers may undertake negative NPV project or invest too many project they company can offer, and most of the project may not benefit for the shareholders. It is called overinvestment of free cash flow since dividends are not paid in that cash flow.
Overall, there are two way can reduce agency costs. First, turn managers to shareholders by give their large proportion of shares (Hawawini & Viallet, 2010). This will be increase the manager's motivation to running the firm day to day business and also mitigate the loss from the conflict between the managers and shareholders. Deft financing can be one way of reduce agency cost and it can avoid the mangers use available free cash to meet their personal pursuits (Jensen ,1986).
2.6 Market timing theory
For a long period of time and under the efficient market hypothesis, there are two competing theories that explain the financing decision: trade-off theory and pecking order theory. However, many studies have challenged these two theories, one of more recently and proposed by Baker & Wurgler (2002) is market timing theory. According to the theory, managers make financing decisions in the light of capital structure conditions (Graham & Harvey, 2001). In other words, managers will issue shares when the stocks are overvalues and tend to issue debt when the stock prices and declines (Baker & Wurgler, 2002). Moreover, Loughran and Ritter (1995) pointed out that ' the long-run relative underperformance of stocks after initial public offering or secondary equity offerings', which is consistent with the market timing theory.
Market timing theory is another theory on capital structure which has been analyzed in recent years. Baker & Wurgler (2002) results that market timing has large and persistent effect on the capital structure of the firm. In this theory, they measure market timing by the external finance-weighted average of historical market-to-book ratios. They find the US leverage of firm is negatively related to this market timing measure. Moreover, a high value could be occurred by this measure when a firm increases its external financing. Indeed, they also find that there could be a significantly positive relationship between capital structure changes and the market timing measure. As a result, they state that "capital structure is the cumulative outcome of past attempts to time the equity market" (Baker & Wurgler, 2002: page). The empirical literatures based on this theory is growing according to the positive evidence made by Baker and Wurgler (2002) for the US. For example, Hovakimian (2006) and Kayhan and Titman (2007) confirm the negative effect of historical market-to-book ratios on US firm's leverage. On the other hand, Mahajan and Tartaroglu (2008) report that a negative relationship between leverage of firma and market-to-book ratio in the G-7 countries, but the equity market timing cannot be a reason to explain this negative relationship. In addition, they also result that the impact of equity market timing attempts on leverage is short lived. This is more consider with dynamic trade-off mode instead of the prediction of the equity market timing hypothesis.
2.7 Empirical Studies on Capital Structure
As has been mentioned above, this research aims to find out evidences to analyze the factors within firms, which can be the determinants of firms' capital structure. As a consequence, to summarize empirical evidences of the major studies will be very helpful to achieve our objective. In this section, we will present 8 empirical studies done in the past years, which main include three countries: United States, United Kingdom and China. As these three countries have been analyzed and studied in recent years, to present these studies could be helpful to understand the area of research.
There are three studies selected in the UK, which includes Bennett and Donnelly (1993), Ozkan (2001) and Bevan and Danbolt (2004). Bennett and Donnelly (1993) used 12 years data which from 1977 to 1988 to analysis the determinants of capital structure. They found firm size, asset structure, non-debt tax shields and firm's past profitability are correlated with capital structure as stated by the theory. However, they reported that a positive relationship between volatility and leverage. Ozkan (2001) apply a dynamic model to exam the relationship between liquidity and leverage, and using panel date that constructed for 390 non-financial firms for the period from 1984 to 1996. The results of the study show that profitability, liquidity and growth opportunities affect a firm's debt ratio negatively and non-debt tax shields affect firm's leverage inversely. Moreover, he only found litter evidence that the relationship between firm size and debt ratio is positive. Bevan and Danbolt (2004) used OLS method and fixed affects method to analysis the determinants of capital structure of 1054 UK companies over the period from 1991 to 1997. The results under the OLS method are consistent with prior literature, but the results under fixed affects method concluded that compared with small firms, larger firms have higher level of long-term debt and short-term debt. They also suggested that there is a negatively relationship between profitability and level pf gearing, and level of growth opportunities only has affect debt ratio little.
In recent years, many empirical researches about capital structure of Chinese firms have been carried out, which support the capital structure theories. A recent study by Chen (2003) that data of 88 Chinese listed companies from 1995 to 2000 has been analyzed by using regression model. It suggests that firm size could affect long-term debt negatively and also a negative relationship between profitability and debt ratio. In addition, they argued that growth opportunities and tangibility are positive related with debt. In another major study, Huang & Song (2005) applied regression model, which contained over 1200 firms' accounting and market data from 1994 to 2003, to analyze the determinates of capital structure of Chinese listed companies. Their empirical results showed that the increase of Chinese firm's leverage will results in the increase of firm size and fixed assets, but will lead to the decrease of profitability, non-debt tax shields and growth opportunities. They also showed that there is insignificant relationship exists between state ownership and firm's capital structure.
Two empirical research has been selected for United States: Ferri and Jones (1979) and Titman and Wessels (1988). On the basis of the data gathered for tow five years period which is firm 1969-1974 and 1971-1976, Ferri and Jones (1979) concluded that industry class directly effect firm's leverage, size and operating leverage are negatively associated with leverage, but a insignificant relationship between variation in income and financial structure. Titman and Wessels (1988) employed factor analytic technique to estimate the determinants of capital structure over a period from 1974 to 1982. The results showed that firm size and uniqueness are negatively related with short-term debt. However, there were not clear relationship between financial leverage and non-debt tax shields, volatility, collateral value of assets, as well as future growth.
This study selects variables as what have been tested in the empirical researches above, to add to the empirical evidences. Thus seven firm characteristics of capital structure are selected in the dissertation. They are firm size, growth opportunities, non-debt tax shields, profitability, liquidity, uniqueness and tangibility.
Firm Size (SIZE)
Firm size plays an important role in determining the capital structure (Booth et al 2001). The majority of studies suggest a positive relationship between size and leverage. Marsh (1982) found that large firms prefer to use long-term debt and small firms prefer short-term debt. Larger firms might take benefits from economies of scale to issue long-term debt, and might even get bargaining power to creditors. Thus, the relationship between cost of issue debt and firm size is negative. However, firm size can also be understood a proxy of the information received by outside investors. Fama & Jensen (1983) points out that compared with small firms, debtors of large companies could always receive more information. Rajan & Zingales (1995) also support this point. Overall, firms with less problem of asymmetric information will have more equity than debt, and have lower level of leverage. On the other hand, the businesses of larger firms are often more diversified and have more stable cash flow. Under the same circumstances, the probability of bankruptcy for larger firms is lower than small ones. Both arguments suggest that size should be positively related with leverage.
Empirical studies such as Rajan & Zingales (1985) analyzed the determinants of capital structure choice by investigating the financing decisions of public firms in the major industrialized countries. They used non-financial firms date for the period 1987-1991. They find significant positive correlation between firm size and leverage except for Germany. A recent study by Antoniou et al (2007) analysed the date from five major countries (United Sates, United Kingdom, France, Germany and Japan) and concluded that the leverage is positively affected the size of the firm.
Growth Opportunities (GROWTH)
Myers (1977) argues that to compare with low-growth firms, high-growth firms may hold more real selections for future investment. If high-growth firms need addition equity financing to implementation such options in the future, a firm has risky debt outstanding may forgo this opportunity, because this kind of investment will make the benefit of adopting the project come to bondholders, which would reduce the wealth of shareholders (Myers, 1977). Thus, in this situation the investment project with a positive net present value cannot be accepted. Since the problem of underinvestment will be greater for firms with greater investment options, Myers predicts that growth opportunities to be negatively related with leverage. Moreover, other authors such as Kim & Sorensen (1986), Smith & Watts (1992), Wald (1999), Rajan & Zingales (1995) and Booth et al. (2001) also support the prediction of negative relationship.
However, as discussed by Myers (1977), Barneal et al (1981), Stohs & Mauer (1996), Barclay & Smith (1996,1999), Michaelas et al (1999) and Ozkan (2000), the growth opportunities affect long-term and short-term debt differently. Michaelas et al (1999) noted this agency problem could be mitigated if the firm issues short term rather than long term debt. Nevertheless, they find the growth opportunities affect short-term debt positively. On the other hand, Bevan & Danbolt (2002) found the growth opportunities affect the total and short term gearing would be positively.
Non-Debt Tax Shields (NDTS)
The non-debt tax shields (NDTS) could be defined as the tax deduction for depreciation and investment tax credits. The non-debt tax shields could be considered as alternatives of the tax shields of debt financing, and a firm has a large amount non-debt tax shields is expected to use less debt (DeAngelo & Masulis, 1980). Therefore, the relation between non-debt tax shields and leverage should be negative.
Both Wald (1999) and Chaplinsly & Niehaus (1993) state the NDTS affect the debt ratio negatively, even they used different ratio to measure NDTS. More recent years, Huang & Song (2005) also found evidence to support this correlation. Some of people such as Bradley et al (1984) argued that a positive relationship between non-debt tax shield and leverage, but most of past studies support for negative relationship.
2.7.4 Profitability (PROFIT)
The theoretical predictions on how the profitability affects the debt ratio could be conflict. On the basis of the pecking order theory, firm prefer internal financing rather then external financing, and retained earnings take first place, follows by debt financing and finally equity financing (Myers, 1984). A much profitable firm could adopt debt financing more, instead of internal financing (Myers & Majluf, 1984). As a result, the pecking order predicts a negative relationship between leverage and profitability. Friends & Lang (1988) used econometric method to test 984 New York Stock Exchange firms from the period 1979 to 1983, and they found evident to support the pecking order hypothesis from US firms. Kester (1986) states profitability affect debt ratio negatively in both the US and Japan manufactures. Moreover, in recent years, Bevan & Danbolt (2002) analyzed the determinates of capital structure of 822 UK firms. The results under regression analysis support the prior literature.
However, in the trade-off theory, the profitability affect debt ratio positively since much profitable firms would prefer debt due to the benefit from tax-shield. Long and Maltiz (1985) find there is a positive relationship between leverage and profitability of firms, but the relationship is not statistically significant. Booth et al (2001) use cross-section and time-series analysis on a sample of 10 developing countries, and suggest that there is always a positive relationship in most countries in their sample.
Liquidity (LIQD)
Mayo (2007: page) points out that liquidity is "the ease with which assets may be quickly converted into cash without the firm's a loss". The liquidity of a firm's asset could usually be measured by the current ratio. According to Ozkan (2001), the liquidity ratios may have more than one effect on the capital structure decision. Firms with higher liquidity ratios may have sufficient capabilities to repay its short-term debt obligations. That is the reason why these firms normally with higher debt ratio. In means that a firm's liquidity is positive related with leverage (Ozkan, 2001). Besides, the liquidity might affect debt ratio negatively, since their investment could be financed by the liquid asset of the company with high liquidity ratio (Ozkan, 2001).
Ozkan (2001) found that liquidity is a negative effect on leverage. Moreover, the potential conflicts between debtholders and shareholders of firms might be the major reason caused this negative effect (Ozkan, 2001). After three years, Deesomsak et al (2004) do the same research with Ozkan, but the date was collected from firms operating in the Asian Pacific region, namely Thailand, Malaysia, Singapore and Australia. They used OLS method to analysis date for the period from 1993 to 2001. Overall they found liquidity to be negatively related to leverage. All of these findings are consistent with the pecking order theory predict which negative relation between liquidity and leverage.
Uniqueness (UNIQ)
In 1984, Titman proposed a model in which the liquidation of firm's decision is causally associated with its bankruptcy status. Consequently, the costs could be transferred to their customers, suppliers and workers by firms through liquidating. It is relevant to their capital structure decisions. 'Customers, workers and suppliers of firms which produce unique or specialized products may suffer relatively high cost in the even that they liquidate'. Furthermore, workers may get job specific techniques, suppliers may suffer the bonding capital, and their customer may difficult to find other substitutes for their relatively unique products (Titman & Wessels, 1988: page). Therefore, the uniqueness affect leverage negatively.
However, some other authors as Ross (1977) suggest that the correlation between uniqueness and leverage should be positive. This result is exactly different with Titman & Wessels's findings. Titman & Wessels (1988) support it by providing evidences. They use the selling expense over sales (SE/S) to measure the relationship between uniqueness and leverage. "This proxy implies that firms with relatively unique products are expected to advertise more and incur costs in promoting and selling their products" (Titman & Wessels (1988: page).
Tangibility (TANG)
It is shown in the empirical researches that there is a negative correlation between tangibility and debt ratio (Jensen & Meckling (1976), Titman & Wessels (1988) and Rajan & Zingales (1995). A recent study by Gaud et al (2005), they analyze the determinates of the capital structure for a panel of 104 Swiss companies listed in the Swiss stock exchange. Dynamic tests are performed for the period 1991 to 2000. It is found that tangibility can affect debt ratio positively. Bevan & Danbolt (2002) also find evidence to support the relationship between them in the UK. One year later, the study of Chen (2003) found tangibility is positively related to leverage. The present study confirms previous findings and contributes additional evidence that suggests tangibility is positively correlated with leverage.
However, some other empirical studies in developing countries found a negative relationship between tangibility and leverage. The first empirical research on the explanatory power of capital structure models in developing countries is conducted by Booth et al (2001). This study has found that tangibility affects debt ratio negatively and this relationship is confirmed by Huang & Song (2005) in China.
2.8 Conclusion
There are three major theories of capital structure have been developed on the basis of the pioneered work of Modigliani and Miller (1958). Trade-off theory is the first theory (Kraus & Lizenberger, 1973), which assumes that firms trade off the debt and equity, and should have optional structure since the market in the real world is far from perfect. In addition, the pecking order theory is the second one (Myers & Majluf, 1984). It concludes that the asymmetrical information should be minimized by adopting financial hierarchy. The third is agency theory (Jensen & Meckling, 1976), it arises when managers and shareholders have interest conflict, and it may lead to many problems such as underinvestment and overinvestment problem. Moreover, the market timing theory as a new theory of capital structure suggested by Baker &Wurgler (2002: page), which states that "the current capital structure is the cumulative outcome of past attempts to time the equity market". The empirical studies covered UK, the US and China three countries documented the empirical evidence about explanatory variables. Therefore, the variables of this study involved firm size, growth opportunities, non-debt tax shied, profitability, liquidity, uniqueness and asset tangibility which reflect the trade off theory, pecking order theory, agency theory and market timing theory.
Chapter 3: Research Methodology
3.1 Introduction
This chapter firstly introduces the method adopted in this study and explains the reasons why we selected this method. Section 2 provides the details on formulation of explanatory variables, and make sure each of variable more suitable in this dissertation. Section 3 and 4 highlights the process of sample section and data collection. The data analysis method presents in Section 5. The last section reports the models used to analysis sample data in this dissertation.
3.2 Quantitative or Qualitative research method
The objective of this study is to determine the relationship between firm characteristics and financial leverage, so it is necessary to choose appropriate research method to achieve the objective. Research methods can be placed into two types: quantitative and qualitative methods. Denzin & Lincoln (2011) pointed out that qualitative research is a multi method in focus on natural setting. This means that this method is more understand the people human behaviors, the most common way to collected date is interviews, questions and surveys.
We are going to adopt quantitative research in the dissertation and the reasons will be provided in the following. Firstly, compared with qualitative research, quantitative research is based on large, representative samples, which are analyzed with statistical tools such STATA and Excel (Spitzlinger, 2010). Secondly, Creswell (2002) argued that one of the advantages of quantitative research is to allow the researchers to test the relationship between independent and dependent variable more objectively than qualitative method. Finally, quantitative research can be applied to test hypotheses, which is more is suitable for this study (Greswell, 2002).
Therefore, overall these three reasons, quantitative research in this case is more appropriate to be applied rather than qualitative method.
3.3 Measurement of Variables
3.3.1 Dependent Variables
There are exists many various measures of capital structure after the seminal paper done by Modigliani & Miller (1958). Generally, there are two main categories of leverage measures: book value of equity and market value of equity. Compared with market values, the book values is more suitable using in this study as dependent variables. And it forces many researchers to measure debt in terms book values rather than market values (Titman & Wessels, 1988). Indeed, the market date is not available for this study. However, Bowman (1980) pointed out that the cross-sectional correlation between market value and book value of debt is huge enough, so the misspecification due to using book value measure can be ignored. Furthermore, most of the researches available in the literature have use book values for their studies, such as Titman & Wessels. Therefore, it is reasonable to use the book values for computing dependent variables.
Benett & Donnelly (1993) and Bevan & Danbolt (2004) suggested that leverage costs of short-term debt maybe not same with the leverage costs of long-term debt. The short-term debt is always used in some emergency situation. For example, when the long-term capital cannot meet the required of firm capital structure, the short-term debt will be used to fulfill the need (Bennett & Donnelly, 1993). Thus, use short-term debt and long-term debt is more powerful to analysis the relationship between short-term debt and long-term debt ratio.
Following the study of Bennett & Donnelly (1993), we use total leverage, long-term leverage and short-term leverage as the dependent variables.
Dependent variables which defined as the book value of total debt that contains both long-term and shore-term liabilities divided by total assets of the firms. Long-term leverage is the book value of long-term debt such as long-term bank loans which can repayable beyond one year divided by the total assets of firms. Short-term leverage is the book value of short-term debt such as bank overdraft can repayable within one year divided by the total assets of the firms.
3.3.2 Independent Variables
In this study seven firm characteristics are suggested as independent variables. They are firm size, growth opportunities, non-debt tax shield, profitability, liquidity, uniqueness and asset tangibility. It is very important to choose appropriate variables, which is closely related to the credibility of the research. Due to this reason, the selection of explanatory variables should be considered the empirical studies from the UK and other countries.
Firm Size (SIZE)
Firm size is measured by taking the natural logarithm of the total assets. The trade-off theory expects a positive relation between leverage and firm size. Large firms is more likely diversified and have more stable cash flow, thus they have low risk for financial distress and less prone to raise bankruptcy cost. Therefore, firm size is become very important sign for a firm when they decided to apply debt and equity. In addition, Marsh (1982) argued that compared with small firms employ short-term loan, large firms are more tend to use long-term loan. The main reason is large firm have benefits of economies of scale to issuing long-term debt and even have bargaining power to creditors. In view of theoretical consideration and empirical evidences, we make the hypotheses as following:
H1. There is a positive relationship between size and leverage.
H2. There is a positive relationship between size and long-term leverage.
H3. There is negative relationship between size and short-term leverage.
Growth Opportunities (GROWTH)
In this study, we follow Rajan & Zingales (1995) used Tobin's Q to measure growth opportunities. Tobin's Q is defined as the ratio of book value of total asset divided by the sum of total liability and the market value of equity. Growth opportunities are viewed as intangible assets of firm. Firms with significant future growth opportunities are likely to face difficulties in raising finance from debt market because intangible assets are not fully collateralisable. Thus, firms with high intangible growth opportunities will use more of equity rather than debt in their capital structure. Therefore, we hypotheses:
H4: There is a negative relationship between growth opportunities and leverage.
Non-Debt Tax Shields (NDTS)
Non-debt tax shield (NDTS) is measured as total annual depreciation divided total assets. Non-debt tax shields such as tax deduction for depreciation and investment tax credits are considered to be the substitutes for tax benefits of debt financing (DeAngelo & Masulis, 1980). Therefore, non-debt tax shields are expected to have negative impact on leverage. This relationship have been supported by many researchers includes Kim & Sorensen (1986), Wald (1999) and Huang & Song (2002). Based on the majority of empirical evidence, we hypotheses:
H6: There is a negative relationship between NDTS and leverage.
Profitability (PROFIT)
Profitability is measured as earnings before interest and taxes scaled divided by book value of assets. The pecking-order theory postulates that firms with higher profits (high internally generated funds) prefer to borrow less because it is easier and more cost effective to finance from internal fund sources. So, as per this theory, there will be a negative relation between leverage and profitability. In contrast, trade-off theory suggests that this relationship would be positive. Since profitable firms are less likely to go bankrupt, and hence can avail more debt at cheaper rates of interest. But most empirical studies find a negative relationship between leverage and profitability in line with the pecking-order theory. We hypothesis:
H7: There is negative relationship between profitability and leverage.
Liquidity (LIQD)
Liquidity is measured as the ratio of current assets by current liabilities. Normally, the previous study showed that two different views about the relationship between liquidity and leverage. However, in this case, we consistent with the pecking order theory, which suggested a negative relationship between liquidity and leverage. It argued that high liquidity firms are more rely on their internal financing and decrease the using of external financing. Based on the previous studies, most of researchers affirm the negative relationship (Friend and Lang, 1988). Consequently, we hypotheses:
H8: There is a negative between liquidity and leverage.
Uniqueness (UNIQ)
Uniqueness is defined as the selling expense dividend by over sales. It suggests that firms will spend more costs on advertising to promote their unique products (Titman & Wessels,1988). Furthermore, as suggested by the trade off theory, firms with relatively unique products will result in lower debt in firm's capital structure (Titman, 1988). We hypothesis:
H9: There is a negative relationship between uniqueness and leverage.
Tangibility (TANG)
Tangibility is measured as a ratio of net fixed assets divided by total assets. Since tangible assets are used as collateral, firms with large amount of fixed assets can borrow on favorable terms by providing the security of these assets to the lenders. Therefore, a high ratio of fixed assets-to-total assets should have a positive impact on firm leverage. Empirical as well as theoretical studies generally predict a positive relation between leverage and asset tangibility. The positive relation between tangibility and leverage is found in Titman & Wessels (1988), Rajan & Zingales (1995), Wald (1999), Chen (2003), Supanvanij (2006), and Akhtar & Oliver (2009). We hypothesis:
H9: There is a positive relationship between tangibility and leverage.
3.4 Sample Selection
The sample includes the Financial Time Stock Exchange 100 index (FTSE 100) at 15th June 2012. According to FTSE website (2012), the FTSE 100 index represents approximately 81% of the UK market capitalization. In other words, firms listed in FTSE that accounts for approximately 81% of the UK market, and it covers wide range of companies from different industries. Therefore, the combination of FTSE 100 could be a represent of most of firms in the UK.
However, not all of 100 firms are selected as a sample in this study. There are two important criteria followed in the process of sample selected. First, the firms in the financial sectors such as banks, mutual fund and insurance companies have different structure of balance sheet from non-financial companies, therefore only non-financial firms will be include from the sample for this study. Secondly, in order to keep the consistence and reliability of date for future research, those firms that have missing observations for any of the variable in the model during the sample period will not included in the date set. Moreover, the time period is chosen from 2005 to 2011 within a 5 years time span. As a result, the 69 non-financial firms over a period of 5 years from 2005 to 2011 will be the final sample.
3.5 Data Collection
The data used in the study are secondary data that collected from Financial Analysis Made Easy (FAME), which provides financial and accounting data for over 3.4 million companies in the UK and Ireland (FAME, 2012). Three types of data are normally used for model estimation. Cross sectional data is data collected by observing many subjects at the same points of the time (Anderson et al, 2008). Time series data is data observed over different time period (Verbeek, 2008). Panel data combine both time series and cross-sectional components. There are many benefits to use panel data in the research even with some problems such as data collection problem, but it is still considers a reasonable method (Baltagi, 2001). Thus, we employed a panel data in this study. We consider that cross-section data set is totally accounted for 69 in one year. Furthermore, the time period is cover 5 years from 2005 to 2011. Finally, the panel data contains 345 observations as a final data.
3.6 Analytical Methods
The pooled ordinary least squares (OLS) are employed as a tool to analysis the data. OLS is defined as statistical technique that uses sample data to estimate the true population relationship between two variables. Pooled OLS method is used panel data to run OLS. It is simple and quick benchmark to which more sophisticated regressions can be compared, even it is subject to many limitations.
3.7 Research Models
In order to investigate the relationship between dependent variables and independent variables, three multiple regression models are employed in this study. The models are as follows.
Model 1
Total Leverage = a +β1SIZE+β2 GROWTH +β3 NTDS +β4 PROFIT + β5 LIQD + β6 UNIQ+β7 TANG+ ε
Model 2
Long-term Leverage = a +β1SIZE+β2 GROWTH +β3 NTDS +β4 PROFIT + β5 LIQD + β6 UNIQ+β7 TANG+ ε
Model 3
Short-term Leverage = a +β1SIZE+β2 GROWTH +β3 NTDS +β4 PROFIT + β5 LIQD + β6 UNIQ+β7 TANG+ ε
Where a is constant, β refers to coefficient number of explanatory variables and e is the error term which is the part of the dependent variable that changes randomly in effect of other possible factors not included in this study.
3.8 Conclusion
This chapter taken in consideration the adopted research approach and concluded that quantitative methods are more suitable in this study. After formation of explanatory variables, the processes of sample selection and data collection have been highlight in the chapter. The pooled OLS method was defined as main analysis method to address the research question. The following chapter will present regression results and give detailed explanation of those results.
Chapter 4: Findings and Analysis
4.1 Introduction
The aim of this chapter is to present the analysis and findings of the influence of various firms' characteristics on capital structure in UK over the period from 2002 to 2011. Therefore, this chapter has been organized in the following ways. The first section will present a descriptive analysis and the regression test results of each explanatory variables. After that, return to the hypothesis posed in the last chapter, we will state test the hypothesis and explain the reasons behind. In the last section, we will get the conclusion for all findings and analysis.
4.2.1 Descriptive Analysis
The study focus on the influence of various firm characteristics on capital structure, and employing a panel data consist of 69 companies which listed in the FTSE 100 over the period from 2008 to 2011. The descriptive statistics for the leverage and explanatory variables of sample firms are reported in Table 1. As can be seen from the table, the overall observation for each dependent and explanatory variable is 345. Besides, the following table exhibits the mean, standard deviation, minimum and maximum values for the dependent and independent variables.
Variables
Obs
Mean
Std. Dev
Min
Max
TLEV
345
0.4994
0.1686
0.0683
0.9966
LTLEV
345
0.2362
0.1380
0.0000
0.5867
STLEV
345
0.2632
0.1172
0.0390
0.6440
SIZE
345
10.8517
2.6128
5.5413
16.0918
GROWTH
345
0.6313
0.1842
0.0992
1.0924
NDTS
345
0.03292
0.05657
0.00000
1.00000
PROFIT
345
0.1509
0.0934
-0.4733
0.4808
LIQD
345
0.9846
0.7189
0.1300
7.5800
UNIQ
345
0.4082
0.5370
-0.8589
7.3520
TANG
345
0.7040
0.1888
0.0023
1.4388
Table 1: Descriptive statistics of Leverage and Independent Variables.
In the above table the dependent variables are TLEV (total leverage), LTLEV (long-term leverage) and STLEV (short-term leverage). The explanatory variables include size (Size), growth opportunities (GROWTH), non-debt tax shield (NDTS), profitability (PROFIT), liquidity (LIQD), uniqueness (UNIQ) and tangibility (TANG).
As Table 1 shows, the firm size is much higher than other explanatory variables. It may be we had applied the natural logarithm form of sales to proxy of size, while other explanatory variables are present in percentage.
4.2.2 Testing for Multicollinearity
There are 7 independent variables in our study, so it is necessary to check that there is no multicollinearity between the variables. Mulicollinearity occurs when two or more independent variables are correlated with each other (DeFusco et al, 2011). There are two major approaches for testing multicollinearity, which involves correlation matrix and variance inflation factor (VIF) test.
The correlation matrix (Table 2) shows that the each of explanatory variable is independent with others when all variables correlated less than 0.8(Katz, 2006). However, there is relatively higher correlation among some of the variables, for example, a positive correlation between LIQD and PROFIT, but it still lower than the recommended value, thus, the problem of multicollinearity does not exist in this study.
Table 2: Correlation Matrix.
The next test for check multicollinearity is tolerance values and VIF test. Walker & Maddan (2008) pointed out that 'tolerance tells how much of the variance of an independent variables does not depend on other independent variables'. If the value of tolerance is greater than 0.20, which indicate that there are no problem with multicollinearity (Tufféry, 2011). On the other hand, the VIF indicates whether a predictor has a strong linear relationship with other predictors. Belsley et al (1980) suggests if VIF values below 10, it appears that multicollinearuty is not a problem. As it is shown in Table 3, all of tolerance values are well above 0.2, and VIF values below 10, therefore, there are both no multicollinearity issues in the model.
Leverage
Tolerance
VIF
SIZE
0.931
1.074
GROWTH
0.862
1.160
NTDS
0.956
1.046
PROFIT
0.770
1.298
LIQD
0.753
1.328
UNIQ
0.884
1.131
TANG
0.804
1.243
Table 3: Tolerance Values and VIF test.
Through checking from the correlation matrix, tolerance values and VIF test, we find that there is no problem of multicollinearity. Furthermore, according to
Ratner (2011), multicilinearity is a data problems, we can assume that there is no data problem in this study. Therefore, we continue to use these data in the following test.
4.2.3 Testing for Heteroscedasticity
In this section, we will test other important assumption of OLS, which is called heteroscedasticity. The reason for Heteroscedasticity happens is due to inconstant term for vaiance of the error in the regression model (Moy & Lee, 2000). In the presence of heteroscedasticity, the OLS estimates remains unbiased, but it becomes inefficient (Verbeek, 2008). Also, heteroscedasticity may give misleading results for t and F statistics (Katos, 2001). Therefore, it is necessary to test whether heteroscedasticity is present.
The Breusch Pagan (BP) test is commonly used test for heteroscedasticity in a liner regression model (Breusch & Pagan, 1979). In BP test, p-value can be used as a measurement to indicates of heteroscedasticity. The p-values in TLEV and LTLEV are zero which less than 0.05. It presents that we have to reject the null hypothesis, and therefore, the problem of heteroscedasticity exist. In order to solve the problem of heteroscedasticity and get the valid results, we run the OLS regression with robust standards errors, and the results are showing in the next section.
Dependent Variables
Chi
Prob> Chi2
TLEV
99.43
0.0000
LTLEV
37.37
0.0000
STLEV
3.52
0.0607
Tables 4.: Results of Breusch and Pagan Tests.
4.3 Regression results
Since our test does not have lagged variables, it does not contain the problems of autocorrelation. Based on our test, the result had met the assumptions of ordinary least square. Therefore, we can use OLS to test the significance of these explanatory variables in our study. The overall regression results with robust standard errors are reported in Table 5-7. The r-square for total debt is approximately 61.48% under the pooled OLS model, which means that the model can explain 61.48% of the leverage ratio. The result had listed in Table 5, where size, growth opportunities and uniqueness are significant and positively related to total leverage. Moreover, profitability and and tangibility are positively related to total leverage and non-debt tax shield and liquidity shows a negative relationship.
Table 5: Overall regression results for Total Leverage. Note: P> |t|-this column shows the 2-tailed p-value used in testing the null hypothesis that the coefficient is 0. In this case, we using significant at 5% level.
The r-square for long-term debt is almost 48% under pooled OLS model. It implies that the model fitness is 48.76%. The result shows except than non-debt tax shields, all of the explanatory variables are highly significant. Moreover, size, growth opportunities, liquidity, uniqueness and tangibility are positively related to long-term leverage among those variables. In addition, non-debt tax shields and profitability negatively related to short-term leverage.
Table 6: Overall regression results for Long-term Leverage.
The r-square for short-term leverage in pooled OLS model is nearly 42% and the result show that the coefficients of growth opportunities and profitability are significant and positively related to short-term leverage. Also, there was a positive correlation between uniqueness and shore-term leverage. Moreover, size, non-debt tax shield, liquidity and tangibility are negatively related with long-term leverage, and among these four variables liquidity and tangibility are significant and negatively related to short-term leverage.
Table 7: Overall regression results for Short-term Leverage.
4.4 Discuss
After discussing the overall regression results applied a pooled OLS model, to discuss the result of each explanatory variable to analyze how they determine the short-term, long-term and total debt ratio. Therefore, in the following we will test hypothesis in the previous chapter to see whether or not hypothesis is accepted, and we will explain the reason behind this. Moreover, the results of this study will compare with the empirical studies in the UK.
4.4.1 Size (SIZE)
The regression results show that size is positively and significantly related to total leverage and long-term leverage, therefore Hypothesis 1 and Hypothesis 2 is accepted. Moreover, we found a negative relationship between size and shore-term leverage, but insignificant, therefore there could not be significant relationship between them, so Hypothesis 3 is rejected. In this research, a 1% increase in firm size will result in an increase in total leverage by 0.0040567%, long-term leverage will increase by 0.0056803% and short-term leverage will decrease by 0.0016237% ceteris paribus.
Dependent Variables
Pooled OLS
TLEV
0.0040567
0.036
LTLEV
0.0056803
0.006
STLEV
-0.0016237
0.400
Table 8: Regression Results for Size.
The positive relationship between firm size and total leverage and long-term leverage shows that the larger the firm sales, the more debt in its capital structure. This finding is consistent with the trade-off theory. According to the trade off theory, firms with larger size have more intensives to be diversified, thus these firms should have lower bankruptcy risk compare with small firms. They may also be able to take advantage of economies of scale in issue long-term debt or borrow more debt as their source of financing (Graham et al, 1998). Therefore, for those of firms listed in the FTSE 100, it is much easy for them borrowing the money from bank and in accessing credit market. The overall results support the trade off theory. Another possible explanation is that individual manages have more control when larger firms have a more dilute ownership, thus they may issue debt to reduce the risk of personal loss resulting from bankruptcy (Friend & Lang, 1988). The result is consistent with the previous research done by Rajan & Zingales (1995) and Wald (1998), which suggested that size was positive related to debt based on the data from developed countries expect Germany.
As shown in the Table 4.6.1, that we had discovered a negative sign between firm size and short-term leverage which is opposite to the measurements of previous two dependent variables. Consider together with the relatively higher long-term debt, we concluded that the firms listed in FTSE 100 prefer use long-term debt rather than short-term debt.
Our results are consistent with the majority empirical studies in UK. Bennet & Donnely (1993) found a positive relationship between size and total leverage as suggested by the trade off theory. On the other hand, the study proposed by Ozkan (2001) which supported the pecking order hypotheses. The study only had shown litter evidence to prove the positive relationship and concluded that firms more rely on debt financing more than equity which implies a negative relationship between leverage and size. The study from Bevan and Danbolt (2004), showed that firm size is positive associated with total debt and long-term debt, but they failed to find evident the relationship between size and short-term debt.
4.4.2 Growth opportunities
We will test another important factor of growth opportunities. Since Hypothesis 4 suggested a negative relationship between growth opportunities and total leverage. However, the association between growth opportunities and total leverage becomes significant and negative in the UK, thus the Hypothesis 4 is rejected. The result of OLS estimation model exhibit there are 1% increase in growth opportunities will results in an increase in total leverage by 0.6456279%.
Dependent Variables
Pooled OLS
TLEV
0.6456279
0.000
LTLEV
0.4783908
0.000
STLEV
0.1672371
0.000
Table 9: Regression Results for Growth Opportunities.
According to the trade off theory, firms holding more tangible assets such as building and machine, which tend to borrow more debt than firms holding more intangible assets such as goodwill, because building and machine can be collateralized (Myers, 1977). Furthermore, agency theory suggests that firms are intent to take advantage from debt holders' wealth (Jensen, 1986). Myers (1977: page) concluded that "firms with greater growth opportunities have more flexibility to invest suboptimal project, and thus expropriate wealth from debt holders to shareholders because of the asset substitution effect". Thus, growth opportunities could affect total leverage negatively.
However, our findings contradict the trade off theory and agency theory. One reason may be that most of firms listed in FTSE 100 possess more tangible assets and less intangible assets, thus they intend to borrow more debt because the tangible assets are easily to collateralize for debt. Another reason may be that high growth UK firms tend to rely on debt financing especially long-term debt to pay for their investment opportunities. As shown in Table 4.6.2, the coefficient of long-term debt (0.4783908) is much higher than the shore-term debt (0.1672371).
The most important reason may be that banks and the equity market all realize the value of growth opportunities and their realization has been expressed in share prices (Chen, 2004). As suggested by the signaling model, compared with low value firms, due to the problem of dead weight costs, firms with high values are able to use high amount of debt financing, and they have low risk to suffer from bankruptcy (Ross, 1977). Therefore, the firms which have optimistic growth prospects are likely to use more leverage. Lang et al (1996) future argued that a negative correlation between growth opportunities and leverage only occurs when the capital market does not recognize firm's growth opportunities. In this study, our sample of firms represents almost 81% of UK market capitalization and all firms are listed in LSE, it indicated that the capital ma
