Capital market

Размещено на http://www.allbest.ru//

Размещено на http://www.allbest.ru//

Content

Introduction

Part one. Literature review

Chapter 1: Capital Structure Theory

1.1 Capital Structure Definition

1.2 Capital Structure Theory

1.3 Trade-Off Theory

1.3.1 Static Trade-Off Theory

1.3.2 Dynamic Trade-Off Theory

1.4 Summary of Capital Structure Theory

Chapter 2: Empirical Research

Part two. Analysis

Chapter 3: Capital Structure Determinants & Hypotheses

3.1 Profitability

3.2 Size

3.3 Asset Tangibility

3.4 Growth Opportunities

3.5 Risk

3.6 Non-Debt-Tax-Shields

3.7 Adjustment Speed

3.8Macro

3.9Repayment

3.10Promos

Chapter 4. Research Methods

4.1 The Sample Data

4.2 Model Specification

4.3 Estimators and Econometric Issues

4.3.1 Trade-Off Model

4.3.1.1 The Estimators

4.3.1.2 Homoskedasticity, Serial Correlation &Endogeneity

Chapter 5. Empirical Results, Discussion and Analysis

5.1 Descriptive Statistics

5.2 Trade-Off Theory

5.2.1 Target Capital Structure

5.2.2 Target vs Observed Capital Structure

5.2.3 Speed Of Adjustment

5.3 Summary of Findings

PART THREE: CONCLUSIONS

Chapter 6: Conclusion and Evaluation

6.1 Conclusions

6.2 Limitations Of Study

6.3 Contributions Of Study And Recommendations

References

Introduction

A vast volume of research papers related to capital structure investigation has been done since Modigliani & Miller had made their scientific work on corporate finance and cost of capital go public in 1958 (Modigliani & Miller(1958)). That moment is supposed to be a start point of capital structure irrelevance theorem. Scientists have been aspiring to create a sole theory to see whether there is an optimal capital structure and to explain firms' financing behavior. Currently there are three theories that explain capital structure and trade-off theory (Modigliani & Miller(1963)) is the most dominant one. The trade-off theory supposes the firm's capital structure choice being a result of trade-off between debt benefits (interest on debt reduces taxation base) and debt costs (direct and indirect bankruptcy costs (Myers (1984))).

There is a great number of empirical studies testing capital structure theories, their explanation force and identifying the determinants needed to be considered when deciding on capital structure choice. However, despite the amount of existing papers the theories are limited to explain the capital structure choice (Graham & Leary(2011)). Nevertheless, noticing the focus on this topic highlights the relevance and importance of this area of study as scientists keep arriving at new views and directions to exercise in future research.

Overall, the recent development has highlighted the importance and relevance of capital structure choice. What factors have impact on capital structure choice? What is the most attractive financing source? Do headquarters strive for optimal capital structure? There are still no exact answers on these questions as it is hard for the sole theory to explain capital structure choice. This research paper drops a light onto yet unexplored area of retail sector by providing a comprehensive analysis of retailers' capital structure.

RESEARCH FOCUS AND OBJECTIVES

Even though this topic is discussed for more than 50 years there are a lot of concerns and the relevance is still quite high.

A great number of papers have tested capital structure on the basis of trade-off theory, so, the chosen topic and research methodology is not new. However, most of them so far were focusing on developed capital markets (US, particularly, (Fischer, et al. (1989); Titman &Wessels(1988); Shyam-Sunder & Myers(1999)), Swiss firms (Gaud et. al.(2005))), cross-country analysis (Jong et. al. (2008); Rajan&Zingales(1995); Alves& Ferreira (2011)), etc. None of the studies has included retail. However, it has been proved that firm's capital structure is influenced by a number of indicators. Some cross-sector research papers have revealed that industry characteristics have impact on it as well. For instance, oil &gas firms rely mostly on equity and bond markets to get access to external sources while retailers rely mostly on equity and leasing. This means that they have smaller leverage level (debt/equity) than other industries, on average, do. This fact proposes capital structure determinants to differ across industries as well as determinants' impact on leverage. Also, this industry is quite cyclical, so, its capital structure is irrelevant to market conjuncture. These are the main reasons to analyze a particular industry.

Retail is seen as a quite attractive industry to study. P&L and balance sheet are more visible and, hence, there are less troubles caused by valuation comparing with companies having business focus on services or manufacturing. Also, retail firms provide more opportunities to compare them due to industry specifics.

One more interesting thing about retail is damping. It means that business decreases the prices so that sales result in negative operating profit. This is done to increase the market share and customer traffic.

As there are almost no research papers on capital structure of retailers there is a gap in existing literature and due to industry specific factors the overall aim is formulated as following:

To identify industry-specific capital structure determinants of retailers and to propose a new determinant to have impact on leverage.

This identification consists of several research tasks. At first, it is vital to find a proof for trade-off theory in retail. Hence, it is:

To find a proof for trade-off theory within retail sector.

The results obtained through the analysis should be compared with the existing literature to show the value of current research in academic area. Thus, we arrive at the second task:

To discuss and to compare theoretical papers towards trade-off theory as well as the results on the capital structure determinants of retailers with the empirical research papers having focus on other industries and cross-country analysis.

Also, due to including a determinant having nothing in common with other papers there is the third task:

To estimate the impact of new determinant on leverage level in case its significance is found.

These research tasks combines provide the necessary basis to achieve the research aim.

DELIMITATIONS

This research paper examines the capital structure determinants of retailers and hence it is focused on those aspects that are relevant for the subject solely. Here we will cover only trade-off theory as it is principally used in empirical studies and hence it gives the opportunity to compare results obtained with a large volume of existing literature. The focus of this degree paper is on the determinants which provide explanation towards the retailers' capital structure. In this regard, there will be no attention given to dividend policy, for instance, and hence the trade-off theory solely will be used to explain the debt amount in the capital structure.

OUTLINE STRUCTURE

The paper is organized in the following way. Part One contains the literature review. Chapter 1 discusses the trade-off theory. Chapter 2 represents the empirical results obtained by researchers. Part Two contains the analysis. Chapter 3 discusses the capital structure determinants in relation to existing empirical literature and provides hypotheses regarding determinants effect on retailers' capital structure. Chapter 4 discusses the research methodology to provide the econometric analysis. This includes econometric estimators discussion and potential problems that may arise. Chapter 5 discusses the obtained results. Part Three contains conclusion and provides recommendations to conduct future research.

Part one. Literature review

This part contains trade-off theory explanation to determine the capital structure determinants later on. Along with the theory empirical results are discussed to the basis of the analysis.

Chapter 1: Capital structure theory

In this section we are about to discuss the main ways of capital structure research development since it has been established as an issue. It would involve a full overview of the theories related to capital structure which had the greatest influence on this area and their impact.

The author would like to start this section by making a review towards the progress having occurred in capital structure theory since Modigliani & Miller made their famous paper go public in 1958 and by presenting the core theories towards capital structure. The amount of surveys and scientific papers on capital structure research is enormous and hence it is out of the scope of this degree work to provide a comprehensive review on the whole impact made by the scientists on capital structure theory. However, the author would like to include the most recent and influential contributions. Initially, we define the capital structure.

1.1 capital Structure definition

As capital structure of the firm is supposed to be the core element in this degree work, so, it is vital to give a definition on what the concept of a firm's capital structure is. By taking a look at the capital structure literature we will find that authors use various definitions. Brealey, Myers, & Marcus(2009), for instance, see the capital structure as the mix of long-term debt and equity financing. On the other hand, as capital structure is related to the way firms provide financing their assets it is not appropriate to take into account long term debt and equitysolely within the capital structure definition, as it is likely for them to issue short term or convertible debt in order to provide financing. There are no doubts that the choice is totally up to both company preferences and the nature of the asset requiring financing. Thus, Welch(2011) challenges the Brealey, Myers, & Marcus statement and suggests instead a ratio of total liabilities to total assets. According to this leverage measure we might say that capital structure consists of financial liabilities and equity. For the literature review purpose and the treatment of capital structure theories it is enough to find a compromise between the two above mentioned and to use the “average”. We say that capital structure consists of long and short term financial debt and equity measured as market capitalization.

1.2 Theory of capital structure

Theory of capital structure has been originated by the paper of Modigliani and Miller (M&M) whose famous article went public in 1958. A great majority of finance academics and businesses are aware of M&M's capital structure irrelevance proposition and some corporate finance textbooks explain cost of capital and capital structure with the M&M paper, for example, Berk&DeMarzo(2011); Hillier, et al. (2008). Basically, the core M&M discovery was that the firm's value and the cost of capital were irrelevant to the choice of financing source, given some assumptions are true. It is also known as the debt irrelevancy theorem.

The conclusions made by M&M resulted in a break with the traditional view on corporate finance. It was a kind of trigger towards numerous debates, lots of criticism and endless volume of articles to go public on this subject. In 1959 Durand appeared to be one of the first to challenge the M&M work. Basically, he criticized the assumptions M&M had made and claimed that in real-world conditions the conclusions they had come to appeared to be false. Durand's statements have initiated the main stream of thought-provoking criticism as many scientists and businessmen highlighted the issues arising from M&M's strong prerequisites that would hardly ever become true within conditions that investors and firms were running business in. Particularly the one stating perfect markets and hence no market imperfections existing is a strong prerequisite as it pays no attention to agency & bankruptcy costs, taxes and supposes all information to be reflected at once in the market and, also, all market players to have equal access to the information. Nevertheless, despite the criticism related to M&M's assumptions and restrictions their paper is still seen as a corporate finance keystone. The thing is that in 1958 their model and affirmations gave to area of corporate finance a tool to systematically provide an analysis towards the factors having impact on deciding what capital structure to choose. Providing perfect markets conditions and hence excluding such determinants as agency & bankruptcy costs, information asymmetry and taxes M&M substantially arrived to the core factors determining capital structure. The prerequisites were then analyzed in a structured way and these actions resulted in extension of theories towards capital structure.

1.3 Trade-off theory

Considering current theory of corporate finance and the views towards M&M's statements related to the optimal capital structure and cost of capital it is not likely to be far from what many academics study today. Nevertheless, it is necessary to find the difference in the factors that cause the company value and average cost of capital to change, where the tax advantage makes a positive impact on the cost of capital and company value by taking up debt financing sources and the moderating effect arising from paying interest on the agency costs caused by the issue of free cash-flow (Jensen(1986)). At the same time, a leverage increase causes the probability growth of bankruptcy and financial distress costs, so, it's vital to find a right balance between debt benefits and the potential debt costs incurred. That is the trade-off theory spirit.

M&M have already dropped a light on the fact that debt influences on cost of capital and company value positively. However, they end up with the following conclusion:

“…with a corporate income tax under which interest is a deductible expense, gains can accrue to stockholders from having debt in the capital structure, even when capital markets are perfect. The gains however are small…”

In their paper M&M stated that the gains from debt financing were almost significant. Nevertheless, in 1963 M&M improved their initial conclusion related to the debt advantage when a firm pays tax on its interests using the corporate tax rate. Hence, they said that the value of using debt as a financial source was higher than it was suggested in the original paper. The value of a levered firm, VL, equals the value of the unlevered one, VU, added the interest tax shield value, фD.

However, M&M do not pay attention to the potential costs caused by debt increase. Meanwhile, according to their additions to the model companies would have the greatest gain by being totally debt-financed. In 1977 Miller stated that bankruptcy cost is quite low compared to the debt benefits and claimed that in case of considering corporate and personal taxes the tax benefit to be high the higher changes would occur in capital structure along with tax rates changes. Miller challenged the papers written by Baxter in 1967 and Kraus &Litzenberger in 1973. Baxter concentrates on the bankruptcy costs and the impact they have on the cost of capital. Baxter mentions that the cost of capital decreases slowly to temper the debt levels but it has a quick increase when the debt-equity ratio is so high that there is a great risk of the firm to face default and to incur the bankruptcy costs both direct and indirect. Direct costs involve legal and administrative costs arising when the firm faces default. Indirect costs are more extensive and it's harder to measure them. They are associated with, for instance, losses on operating sales due to the lack of customer convenience, loss of loan availability given by suppliers, incapacity to prevent a work force flow, etc. Having considered these costs we might state that they provide a convex function related to the cost of capital along with variety of debt-equity ratios which is close to the existing views and to the current cost of capital conception known as WACC (Weighted Average Cost of Capital).

1.3.1 Static trade-off theory

Basically, trade-off model deals with the costs and gains arising from debt issuance. It is often associated with Krauss &Litzenberger paper that went public in 1973. They provided a single-period valuation model which accounted for both the tax advantage value arising from debt issuance and potential bankruptcy costs. According to Kraus &Litzenberger we might divide the levered company value into three categories: the value of unlevered firm; tax advantage gains and after tax bankruptcy costs. With this researchers suggest levered companies to increase their value by finding the right balance between two ending components (tax advantages and bankruptcy costs). It leads us to the conclusion that a firm has an optimal capital structure that is able to increase its value by debt issuance until it reaches the critical point when tax advantage from marginal debt issuance is counterweighted by rise in bankruptcy costs.

According to Kraus &Litzenberger and, also, Myers(1984) the model assumes that firms are constantly to have an optimal debt-equity ratio. This permanent equilibrium states that the only possible way for the companies to change their capital structure is sudden shock that makes their equity market values drop. In real-market conditions it is quite unlikely as debt or equity issuance or repurchasing inevitably generates transaction costs and hence there is higher probability that firms would rarely adjust their capital structure reflecting the transaction costs incurred. Original Kraus &Litzenberger model doesn't account for these transaction costs. On the other hand, Myers was aware of such costs and he implied them as a possible explanation why firms didn't change their capital structure as frequently as the static trade-off model predicted.

As it was mentioned before, the model is a single-period one, so, it deals with a firm's capital structure paying no attention to the other periods. This is the core disadvantage of this approach. Capital structure in each period is inevitably correlated with the one in preceding periods and firms are likely to account for future expectations when making decisions on debt-equity ratio.

There are numerous empirical researches of capital structure determinants been performed through the past two-three decades. They have resulted in various findings towards relations between a firm's leverage ratio and its certain characteristics. One of the most coherent conclusions in empirical research towards capital structure is that lower leverage gives company higher profitability (Fama& French (2002)). Hence, the result contradicts with the static trade-off theory as it states that higher profitability requires an increase in leverage ratio to gain from a tax shield arising from higher interest. Moreover, the static nature of this theory concludes that firms always have an optimal capital structure immediately adjusted by sudden changes which, also, doesn't fit real-market conditions (Flannery &Rangan(2006)).

The contradictions between static trade-off theory and the empirical outcomes along with the static models inability to provide company analysis that runs its business in multi-period conditions have made researchers to carry out dynamic trade-off theories that claim promising results.

1.3.2 Dynamic trade-off theory

Dynamic trade-off theory is associated with Fischer, Heinkel and Zechner (FHZ) paper written in 1989. This research group was one of the first to challenge a static trade-off and to create a model that assumed companies deviating from their optimal capital structure. Number of incremental discontents related to the static trade-off theory has brought to numerous contributions to dynamic model since FHZ work had been published. Dynamic trade-off model sees firms' capital structure as an ongoing decision taking into account a compromise between the tax gains from debt and potential bankruptcy costs and, also, restructuring costs and investment decisions. Unlike static trade-off theory the dynamic one supposes adjusting capital structure costs to have firms deviating from the optimal capital structure for longer time.

FHZ model suggests firms having a range of optimal debt-to-equity ratio to let it vary depending on circumstances. The core idea is that capital structure adjustments are not performed immediately due to shocks in firms' asset values. Instead, they let their capital structure fluctuate within a range. The reason is that adjustment costs within this range surpass the gains of doing so. Only at firm's specific lower and upper capital structures is it worth adjusting. In the figure below there is the model's fundamental idea.

Firms let their debt-equity ratio fluctuate within the boundaries of and as the costs of capital structure adjustment within these lines would surpass the gains. If the firm's debt-equity ratio is lower than the it will make adjustments as the gains from debt issuance exceed the adjustment ones. At the upper limit the risk of potential bankruptcy rises and those costs exceed the ones linked with capital structure adjustment.

The FHZ model suggests an engaging multi-period dynamic prospect to the financial decisions made in firms. However, it doesn't account much for implications necessary for empirical research. In 2007 Strebulaev created a dynamic trade-off model that included costs related to capital structure adjustment and generated coherent within observed capital structures results. Strebulaev pays a lot of attention to the fact of diversity between firms' debt-equity ratio at refinancing point and their actual ratio when he collected the data. Like FHZ, Strebulaev stated that because of adjustment costs firms are to keep their optimal debt-equity ratio only by reaching individual refinancing points. When analyzing a cross-section of firms it is quite unlikely for them to be at their refinancing points during data mining and, moreover, firms stay away from their optimal debt-to-equity ratio differently and hence within static models it makes cross-section analysis imprecise which may result in wrong conclusions. For instance, within static models negative correlation between leverage and profitability has been seen by majority of researchers as a pecking order theory proof and hence they rejected the trade-off one. However, in Strebulaev's paper we find how this correlation can be explained by using a dynamic trade-off model. When firms face a profitability increase that influences on firm's value positively the future profitability expectations improve. However, due to adjustment costs firms do not adjust at once. As a result, the leverage ratio goes up.

FHZ and Strebulaev models show good specifics. On the other hand, they solely account for costs related to capital structure adjustments and, as a strong assumption, suppose firms' investment decisions being independent on their financing choice and exogenous. Unlike, firms are to account for future investment decisions while considering present financing choices.

Other scientists have created alternative dynamic models including more factors. In 2005 Hennessy & Whited (H&W) suggested one combining firms' financing and investment decisions continuously. It assumed firms taking up both financing and investment decisions each period mutually which makes us conclude firms always to be at a refinancing or restructuring point. H&W made a controversial conclusion stating that firms have no target debt-equity ratio. However, leverage is supposed to be a result of a firm's business in the past and its future expectations. Also, in 2001 Graham and Harvey mentioned that 71 % of the surveyed firms had a tight or flexible target debt ratio. Recent empirical and theoretical researches related to a dynamic trade-off theory had a particular focus on firms' target debt-equity ratio and the speed companies made adjustments towards their target capital structure.

In 2007 Titman and Tsyplakov created a model that accounted for many factors of H&W model and found out that firms tended not to have a target leverage to which they made adjustments. The speed of doing so is not high and it depends on factors like costs of financial distress. Also, agency costs do influence on firms' adjustment inducement. Both Strebulaev& Whited (2011) and DeAngelo et al.(2011) as well stated slow speed of adjusting and concluded that firms are willing to use debt conservatively to be financially flexible. DeAngelo et al. said that firms prefer using use temporary debt to fund raise for investment projects and, also, to keep using this debt in the future they maintain below their debt capacity. In 2012 S&W included financial flexibility into their model and concluded that capital structure adjusting opportunity is valued by firms and due to this the optimal debt-equity ratio is at the lower level as well as the speed of adjusting. Accounting for financial flexibility within dynamic models is pertinent as there is a strong evidence found by Bancel&Mittoo(2004) and Graham & Harvey (2001) who found out that chief executive officers treated financial flexibility as an important factor (91 and 59 percent, respectively).

The observed dynamic trade-off models have much in common with conclusions of empirical researches made earlier. On the other hand, the results obtained and conclusions are based on synthetic data that tends to carry out results more related to the theory. The framework for empirical research interpretation is improved by the theory elaboration underlying. Currently, there is an increasing number of empirical researches having adjustment speed determination as a core purpose.

1.4 Summary of capital structure theory

This chapter discusses the trade-off theory as the most dominant theory of capital structure. It provides views and explanations on capital structure choice. The next chapter elaborates on the empirical results of research papers and the support they offer to the trade-off theory.

Chapter 2. Empirical research

The purpose of empirical research is to define what capital structure is used by the firms. Initial papers were to identify the factors having impact on debt-equity ratio and then their correlation was justified according to the capital structure theories (Rajan&Zingales, 1995; Harris &Raviv(1991); Frank &Goyal(2009); Titman &Wessels(1988)). The great majority of those tests was performed within static prerequisites, so, that's the reason why the conclusions made hardly were true. The core purpose of these papers was to interpret the firm's debt-equity ratio by the determinants that theoretically might influence on it. Some of them are tax gains, agency & bankruptcy costs, etc.

Empirical researches performed recently account for not only the debt-equity ratio but, also, the changes occurring. They are based on the theories within dynamic framework and the concept stating that some factors can make firms get sidetracked from the optimal level of leverage. Leaning on the dynamic trade-off theory implications researchers use target adjustment models to identify how fast firms make adjustments towards their optimal leverage level.

The table below considers the results for speed of adjustment (SOA) and other commonly used determinants of capital structure. Though the majority of researchers performed their analysis within the US firms sample it still fluctuates greatly(the range is 5-34 %). The reason might be the specification of the models.

Table 1. Summary of empirical papers (own contribution).

Paper	Size	Growth	Profitability	Non-debt tax shield	Volatility	Tangibility	SOA
Titiman&Wessels (1988)	+	-	-	-	-	-
Harris &Raviv (1991)	+	+	-	+	-	+
Rajan&Zingales (1995)	+	-	-			+
Shyam-Sunder & Myers (1999)		-	-			+	30%
Baker &Wurgler (2002)	+	-	-			+	22-25%
Fama& French (2002)	+	+/-	-	-	-	+	7-17%
Gaud, et. al. (2005)	+	-	-			+	27.3%
Flannery &Rangan (2006)	+	0	-	-		+	34.4%
Jong, et al. (2008)	+	-	-	+	+
Frank &Goyal (2009)	+	-	-			+
Alves& Ferreira (2011)	+	+/-	-			+/-
Hovakimian& Li (2011)	+	-	-	+		+	5-8%
Kokoreva&Belozerov (2014)	-	-	+		+	-

Rajan&Zingales; Jong, et al. and Alves& Ferreira have considered countries' specific capital structure determinants and their analysis is based on cross-section sample. Rajan&Zingales studied G-7 countries and came to the conclusion that the determinants and capital structure correlation is basically the same but, however, Germany has shown the negative one for size. The authors linked it with more strongly marked bank support towards loaning. Though Jong, et al. have approved their results of pooled sample they found out that capital structure determinants varies depending on the country. They concluded that stock markets development has a strong influence on the firm's leverage level. As well, Alves& Ferreira faced the difference in capital structure determinants within the sample of countries. However, they treated shareholders' rights as the core reason of this difference. Gaud, et al. were to deal with Swiss companies sample and they made similar to US-investigating researchers conclusions and got SOA being equal to 27.3%.

In 2014 M. Kokoreva along with I. Beloserov investigated capital structure of BRICS firms and tested whether dynamic trade-off theory was true for them or not. The paper states that there is a range of an optimal capital structure it volatiles within due to transaction costs. Also, the determinants appeared to have the influence trends common with the other scientific papers. The third statement says that capital structure tends not to adjust to the optimal level fully.

The core representatives of another approach are Fama& French. They are dealing with assumptions towards dividend police and debt within capital structure theories. The results they achieved do not support any theory and thus Fama& French state the following:

“In sum, we identify one scar on the trade-off model (The negative relation between leverage and profitability), one deep wound on the pecking order (the large equity issues of small low-leverage growth companies), and one area of conflict (the mean reversion of leverage) on which the data speak softly.”

Fama& French conclusion considers two important issues. These are negative correlation of profitability and debt-equity ratio; and how we should treat with the adjustment speed (mean reversion according to Fama& French). We have already discussed the first issue above and stated that this correlation is the result of numerous factors like firm's need to remain being financially flexible or adjustment costs. Empirical researches often rely on static trade-off and perform the analysis within the cross-section. One of their core assumptions is that firms are currently at the refinancing point.

Secondly, 100% adjustment speed tells us that there's no time lag between leverage changes and adjustments made while 0% one points out that target capital structure is not the number-one factor for the firm. Adjustment speed at lower levels like in Fama& French and Hovakimian& Li research papers are seen as a contradiction with the trade-off theory. On the other hand, DeAngelo, et. al. and Strebulaev& Whited (2012) state the low adjustment speed to be an aspiration to maintaining being financially flexible. So, as we can see there are four core factors to be coherent within the empirical researches: asset tangibility and size show positive impact while growth opportunities and profitability remain having negative one.

Part two. Analysis

retail sector capital

The previous chapters discussed trade-off theory and empirical results obtained from a great number of scientific papers towards the capital structure determinants. They include the determinants that are commonly used to explain the capital structure. Current part contains analysis. Chapter 4 discusses the capital structure determinants and provides hypotheses based on theory and the empirical results highlighted. Chapter 5 contains the model specification along with research methodology. Finally, chapter 6 discusses the obtained results from the trade-off model estimation. Also, it compares them with the existing literature represented in the second part.

Chapter 3. Capital structure determinants & hypotheses

According to the research paper considered above we are able to sum up common factors having most impact on firm's capital structure in retail sector. In this part of the work we are going to exercise them and to suggest hypotheses towards the influence they have on the listed firm's debt-equity ratio.

3.1 Profitability

Profitability tends to be the key determinant observed almost in each empirical research paper towards firm's capital structure. The trade-off concept sees the higher level of profitability as a reason of debt increase. There are two reasons for that. First of all, as the cost of debt is less than the cost of equity and hence debt financing is less risky highly profitable firms are less likely to face bankruptcy. Secondly, higher level of profitability lets the firms increase their levels of interest tax shields by raising the debt-equity ratio and hence paying more interest to the debt holders. Also, according to Jensen(1986) free cash flows are to grow within an increase in profitability. However, the dynamic trade-off concept assumption that firms wouldn't adjust their leverage level immediately due to adjustment costs and the fact that being at the refinancing point during scanning is quite unlikely for the firm result in the negative correlation between profitability and debt-equity ratio because of the static frameworks of the analysis.

As observed in empirical research profitability tends to be negatively correlated with the leverage level. Therefore we estimate the following hypothesis:

H1: We expect a negative correlation between profitability and leverage level within retail sector firms.

In empirical papers profitability is measured differently. Some of them use the ratio of operating profit to total assets (Frank &Goyal(2003); Titman &Wessels(1988)), others hold on to EBIT (Flannery &Rangan(2006); Fama& French (2002)) or EBITDA to total assets (Alves& Ferreira (2011); Rajan&Zingales(1995)). In this research EBIT-total assets ratio will be examined. The reason is that EBIT is supposed to be a better basis compared to EBITDA as we might face D&A treated legally differently through the countries.

3.2 Size

The size is used frequently in empirical researches as well and gives coherent results within capital structure correlation. Frank &Goyal stated that large firms' cash flows tend to be less volatile and are less likely to face a default as they are usually more diversified. Thus, the trade-off concept sees the size to be positively correlated with the debt-equity ratio.

Despite Rajan&Zingales found a negative size influence on leverage in Germany the other research papers tend to have quite opposite results. So, our next hypothesis might be said as:

H2: Large firms rely on debt financing more than on equity.

Scientists usually measure the size as a natural logarithm of total assets or revenue. Here we want to take up the first one.

3.3 Asset tangibility

The core idea here is that tangible assets are seen as a collateral security for the investors. According to Titman &Wessels(1988) as the cost of debt is reduced due to the risk decrease it is possible to focus more on debt financing sources along with lower risk of facing bankruptcy. Hence, firms having more tangible assets in their asset structure would be likely to raise debt. Moreover, debtholders security is based on particular assets, so, investors face difficulties in providing asset replacement and hence there is a decrease in agency costs between shareholders and debtholders.

So, having analyzed empirical papers we arrive to the statement of the next hypothesis:

H3: Retailers having more tangibles in their asset structure tend to be more levered.

The way to measure the tangibility varies across the papers as well. We determine the tangibility like the ratio of PP&E and total assets.

3.4 Growth opportunities

As well as the asset tangibility, growth opportunities are taken into consideration within the similar reasons, however, the correlation turns out to be negative. In 1977 Myers was the first one to consider the capital structure dependence on growth opportunities. As he concluded, when a firm has various growth opportunities to follow potential investors face difficulty in valuing them or determining which ones would be accepted by the top-management.

The more growth opportunities the company has the more likely it is to face bankruptcy or financial distress as those opportunities have less value, or not at all, for the other firms than for the initial one. Thus, in 1988 Titman &Wessels stated the negative correlation between the growth opportunities and the leverage level considering the trade-off theory. Moreover, the EBIT tends to be lower within firms having a lot of investment opportunities, so, it is hardly possible to apply tax shield associated with the interest payments fully. More than that, it's highly important for the firms to have lots of growth opportunities to be flexible within their finances and hence they are less levered and the optimal debt-equity ratio is lower as well.

H4: Retailers would not be likely to use debt financing when having more opportunities to grow.

In our analysis we are going to use the ratio of market value to book value of equity as the market reflects positively when it is released for the firm to grow, so, the market value exceeds the book one.

3.5 Risk

According to the trade-off theory, risky firms tend to use less debt as the bankruptcy costs arise because of the volatility in earnings. Moreover, according to Frank &Goyal risk reduces the gains from the debt interest tax shields which results from volatility and leads to reducing the amount of debt financing.

H5: Firms that are prone to a high volatility level are less likely to use debt and hence they have lower debt-equity ratios.

Actually, researchers haven't yet come to a common decision on how to measure the risk. However, as it was mentioned we determine risk as volatility in earnings and hence the EBIT standard deviation is to define the risk level. Also, it is often found in scientific papers.

3.6 Non-Debt tax shields

In 1980 DeAngelo and Masulis wrote in their paper that there is no reason for the firms to gain from debt interest when having non-debt tax shields being a substitute to it. Hence, according to the trade-off theory such firms will be less levered.

H6: Using a large amount of non-debt tax shields firms reduce debt share in their financial structure.

Here those shields are going to be measured with D&A as it reduces the taxable basis and thus substitutes the interest payments. As the majority of the researchers do, we will use the ratio of D&A to total assets.

3.7 Adjustment Speed

In 2013 VusaniMoyo&HendrikWolmarans& Leon Brьmmercame to a conclusion stating that capital structure adjustment speed is highly affected by institutional differences across industries. The adjustment speed is reduced with the adjustment costs to arise due to the informational asymmetry, capital markets ease of access and financial constraints having positive influence on the costs. Adjustment speed is increased by using the debt covenants or higher level of the bankruptcy costs and tax rates as those factors provide the firms with the gains from moving towards target capital structure. Also, relying on bank lending implies using debt covenants as they make the firm face the penalty in case they are exceeded and hence the adjustment speed rises up. This results in a conclusion stating that unless overleveraged firms increase their adjustment speed they will face the penalty and the loan rating thus will decrease along with the growth in bankruptcy costs.

This way we get to the last hypothesis:

H7:Overleveraged firms have higher adjustment speed towards target capital structure in comparison with the ones having a little share of debt in their financial structure.

3.8macro

In the introduction we have stated that retailers' capital structure is not affected by market conjuncture as it is a cyclical industry (EY analyst opinion). Moreover, this might be true not for the particular country or region but for the whole industry. Thus, we arrive at the eighth hypothesis stating:

H8: Retailers' capital structure is irrelevant to market conjuncture and the region business is operating in.

As a proxy, we make a dummy variable to state a region. It will be 1 for the USA and 0 for others. Also, GDP growth will be used to estimate market conjuncture influence on capital structure.

3.9 Repayment

In investment banking business there is a ratio commonly used to estimate business willingness to borrow funds. This is Debt/EBITDA ratio. EBITDA here stands for a cash flow. Thus, the more it is, the higher opportunities there are to get a loan in a bank or private investor as it is easier to make interest payments and provide debt repayment. Hence

H9: Debt/EBITDA should be negatively correlated with leverage indicating opportunities to make repayments.

3.10 Promos

Thelast hypothesis is linked with damping. As we have already discussed in the introduction retailers tend to decrease prices to get a positive effect onto market share and SKU (Stock Keeping Unit) sales increase. Promos are closely connected with advertising costs. As usual, a successive advertising campaign (promo) level up the business reputation (goodwill). Also, promos are usually funded with debt as cash flows occur being negative. So, the leverage goes up (EY analyst opinion).

H10: Promos increase leverage level due to debt financing.

As a proxy we will choose ratio goodwill/EBITDA as goodwill increase reflects a good promo and EBITDA shows damping effect.

The next chapter discusses the research methodology used to obtain valid results.

Chapter 4. Research methods

This chapter contains the framework to provide the analysis of retailers' capital structure. It describes the sample and chosen search strategy and, also, data set construction. The trade-off model specification is discussed together with the econometric issues that may arise.

4.1 The sample data

The sample consists of public firms listed in the USA, Canada and Eurozone dating 2010-2015 (but for some firms 2015 data is unavailable yet as their financial year still goes on). Here we deal with retail companies solely according to the paper topic. The sub-sectors are food retail and fashion. The data is collected from the firms' financial statements and databases (Bloomberg, Reuters, etc.). As far as the model is a dynamic one which means it operates with lagged variables it is necessary to have at least two on-going time periods for each firm. In our sample six time periods are used. Finally, there are 60 firms analyzed. To avoid the wrong data interpretation it's necessary to exclude extreme values from the sample. All in all, there are 251 observations currently left due to extreme and inadequate meanings.

The sample itself is represented as an unbalanced panel which means that for some firms there are no full six-years observations. There are three core arguments for using the panel which are particularly relevant for the following topic of research:

It is possible to investigate both individual and time factors within using the panel.

We are able to perform the dynamic capital structure analysis needed for this paper.

The last but not least, we can take the company heterogeneity into consideration.

To the author's mind, to be able to analyze data in a dynamic way and to consider the heterogeneity is quite relevant for the area the paper refers to. Moreover, ignoring the heterogeneity will result in the estimates insignificance obtained through the analysis.

The fact that the sample is an unbalanced panel makes us dealing with some challenges but, however, the advantages outbalance these complexities. It contains a lot more observations and decreases the possible survivorship offset arising when dealing with balanced panel.

4.2 Model specification

The specification is based on the dynamic trade-off theory which is used to determine whether it is relevant for retail companies' capital structure choice decisions or not.

As we have already mentioned in both theoretical and empirical literature review as firms run their business in dynamic conditions their decisions towards capital structure choice should as well be taken into consideration within dynamic frameworks. It's known that when the firm makes the adjustments towards its leverage it incurs costs, so, consequently it doesn't happen at once but does by degrees. To grasp this consistent adjustment we use a model of partial adjustment created in 2006 by Flannery &Rangan.

With this we are able to consider the changes in capital structure within two periods. There is an adjustment speed on the right side of the equation standing at the first place ( ). It tells us how fast the firms make the adjustments towards their leverage level having the previous period as a basis. Speaking about target capital structure we may define it in the following way:



Here stands for a vector representing the determinants. If we integrate it into the previous equation we will get the model which is quite similar to the one existing in Gaud, et al.(2005) paper:

Due to the fact that the model may have missed time effects and time-invariant firm heterogeneity we treat them by way of and , respectively. Time-invariant firm heterogeneity supervenes from successive differences arising between firms analyzed within the regression framework, however, it is not seized by the interpretative variables (Verbeek(2008)). Both different investors' risk profiles and managerial preferences may be considered as factors resulting in successive capital structure differences between the firms.

To provide the full specification of the dynamic trade-off theory model it is better to replace the vector with the list of the interpretative variables:

In the next section we are going to describe the econometrics to be used and the estimators' efficiency.

4.3 The estimators and econometric issues

Using the panel data gives us many advantages in terms of econometric analysis. On the other hand, in case we keep the same cross-section to measure but within different points in time we can no longer claim that there is independence within individual observations. To take the unobservable effects into consideration we need to include several estimators into the model specification. To estimate the trade-off theory within the dynamic linear regression it is necessary to use more complex estimators like GMM, for instance.

4.3.1 Trade-off model

To provide the dynamic analysis of this model we use the leverage ratio regression in a t-time period onto a set of estimators which include the (t-1)-time period leverage ratio. The required strict exogeneity assumption is not true for this model as there is a lagged dependent variable in a set of estimators present. In 2013 Flannery & Hankins showed that performing fixed effects or a standard OLS method would cause biased estimations within the dynamic set. To be more accurate, there will be downward/upward bias within the lagged leverage ratio which will result in an upward/downward bias of the adjustment speed estimation for fixed effect and OLS, respectively.

Having Flannery & Hankins findings as a basis we can say that GMM estimator suits here well as it gives possibility to handle the endogeneity and, moreover, it is applicable for the unbalanced panels. Finally, it is often used to estimate the trade-off model. GMM goes along with the objective to get such estimates that will be possible to be compared with the ones obtained in the papers conducting other industries. The GMM estimator was invented in 1991 by Arellano & Bond and it is objected to deal with the panels having little time observations but with numerous cross-sections. Also, we have taken into account the limited sample bias that is the result of limited samples having several instruments. Hence, we use the first-difference instrumental variable estimator designed by Anderson & Hsiao. Also, based on the intuition we include the opposite-sign biases into the lagged leverage ratio estimation provided by FE and OLS regressions.

As we have covered the FE & OLS procedures and assumptions we will talk about GMM and Anderson & Hsiao instrumental variables estimators in the rest of this section.

4.3.1.1 The estimators

Both GMM and Anderson & Hsiao IV estimators are based upon the instrumental variables and first differences. Anderson & Hsiao have suggested using as instruments either lagged differences or levels dependent variables as instruments. The first-difference equation is constructed the following way:

The estimation will be performed using and as the instruments for . On the other hand, we sacrifice observations using the instruments. We are not able to use before the time period equals 3 and speaking about the lagged difference it is unavailable until the time reaches 4. The loss of some time periods results in a significant loss in observations as corporate finance panel are quite often constructed having many cross-sections but little time-series.

Though the Anderson & Hsiao instrumental variable estimator might be consistent it doesn't consider all the lagged instruments. Thus, in 1991 Arellano & Bond created a GMM estimator based on difference that allows to apply all the lagged values as instruments. For instance, are the instruments for . However, this estimator has some disadvantages like limited sample bias and weak instruments. In addition, in 1998 Blundell & Bond created a GMM estimator that also used lagged differences estimating the equation. For instance, are the instruments for . It should be mentioned that the majority of researchers suppose the GMM estimator to be the best one to use within the unbalanced panels with some variables to suffer from the lack of exogeneity. Moreover, homoscedasticity is not required here.