Investigating the Leverage Composition of Pakistani Firms Through Their Determinants

To have an ideal mix of debt and equity in a balance sheet of an entity is till to date a very complicated issue for managers as there is no such rule to predict an optimal capital structure. An in-depth understanding is required for the corporate culture, the degree of the development of the capital market and the economy in which the firms operate. This study seeks to investigate the leverage composition of Pakistani corporations through their determinants. Fixed effect regression is used to show the relationship of determinants of capital structure on leverage corporations listed on Karachi Stock Exchange (KSE) for the period of 2006 to 2013. The results suggest that agency cost, growth, age, and size are significantly and negatively associated with the capital structure of Pakistan firms, however, collateral value of asset is significantly but positively associated with the capital structure of the firm. On the other hand, free cash flows, non debt tax shield, profitability, business risk and bankruptcy cost are not significantly associated with leverage composition of the firms and are against the signaling theory and peaking order theory. The key importance of this study is that no prior research was done for determinants like agency cost, free cash flows, bankruptcy cost and age as determinants of capital structure for Pakistani firms among other determinants. Further, this study does not confine to a particular sector rather it covers all companies listed by Karachi Stock Exchange.


Introduction
Capital structure irrelevance theory was introduced by Modigliani and Miller (1958) and since then arguments have progressed for the leverage decision of the firm. MM theory suggests that the value of the firm is independent of its capital structure under certain assumptions. Even though this theory is based on unrealistic assumptions, there are a number of variables that narrate the value to the firm and often identify as determinants of the capital structure e.g. agency cost, collateral value of asset, growth, free cash flows, and age of the firm, business risk, bankruptcy risk, profitability and non debt tax shield. Hence, the main purpose of the firm is to maximize the wealth of the stockholders by evaluating a suitable finance mix. Combination of debt and equity capital becomes the most controversial corporate issue over the past four decades. The capital structure decisions directly influence the market value of the firms and the cost of the firm.
How to have an ideal mix of debt and equity in a balance sheet of an entity is till to date a very complicated issue as there is no such rule to predict an optimal capital structure. An in-depth understanding is required for the corporate culture, the degree of the development of the capital market and the economy in which the firms operate. Firms can only achieve their objectives through skillful intellectual managers and the management can perform better without thinking about finance shortage or finance mix.
So soon after the arrival of (Modigliani & Miller, 1958), the western world started financing their corporations without bothering the mode of finance i.e. debt or equity and forgot about the equity debt mix and just focused on growth and achievement of commercial business objectives. Companies achieved local and international remarkable growth. Due to rapid growth historic numbers soon became meaningless. Corporations started focusing more about their market values and future fund flows rather than the book value of assets. Financial ratios are also designed on the market values of the company. Human capital also becomes an important part of the company.
Pakistan is one of the developing countries with lot of issues involving unstable micro and macroeconomic situations, political crisis, social behavior, geographical structure, complicated tax mechanism and non robust legal system. It is equally important to find out those factors that influence a firm's capital structure choice. However, in case of developing economies, inadequate literature is available regarding capital structure of corporations in Pakistan. Eldomiaty (2008) mentioned that because of the insufficient information problems, the capital markets of the developing economies are not efficient enough to compare with developed market. Therefore, the outcomes of the developed countries cannot be generalized with the developing countries like Pakistan where political risk is very high, foreign currency fluctuations are very frequent, business risk is lofty, and the capital market is in developing phase. Most of the debate in the country on low investment ratios has been centered around factors such as infrastructure, law and order, skill shortages and bureaucratic hassles (Hussain, 2006).
The key importance of this study is that no prior research was done for determinants like agency cost, collateral value of assets, free cash flows, bankruptcy cost and age as determinants of capital structure for Pakistani firms. Further, this study does not confine to a particular sector rather it covers all companies listed by Karachi Stock Exchange. Therefore, this study shows an in depth analysis of determinants of capital structure of Pakistani firms. Moreover, this study is very functional for the managers of the corporations and provides guidelines for efficient use of the determinants of capital structure in order to maximize firm performance. Government authorities, taxation bodies and policy makers can also be benefited from the findings of the study. Hence, this study also served the purpose of a rich contribution in the existing literature regarding the determinants of capital structure of firms in Pakistan.

Literature Review
Theories of corporate structure draw a closer attention in the world of finance when Modigliani and Miller (1958) presented capital structure irrelevance principle. This principle states that in a perfect market, whatever the mode of finance the firm uses, it is irrelevance to the firm's value. And since then, many other theories were presented with the help of basic idea provided by Modigliani and Miller (1958) like trade off theory, pecking order theory, OLI theory, signaling theory, etc. These theories highlight different determinants of capital structure and draw attention to their importance in the perspective of capital structure.
Pecking order theory is based on asymmetric information of the firm. In this theory, Myers (1984) explained that how a company prioritized its financing decisions.
The main theme of the theory is that the firm takes finances from easier sources first i.e.
internal finance. Pecking order theory also suggests that there is an inverse relationship between profitability and leverage. Modigliani and Miller (1958)

Determinants of Capital Structure -Around the World
Al Amri and Al Ani (2015) examined determinants of capital structure for three sectors (food, construction and chemical) of Omani industrial companies for the period of 2008 -2012. They found that there is a significant and positive relationship between risk and tangibility and leverage and there is a significant but negative relationship between growth rate and profitability and leverage, while there is no association with size. Baltacı and Ayaydın (2014) explored Turkish banking sector and found that capital structure is positively and significantly related with size, industry leverage and GDP growth. They further find that capital structure is significantly but inversely related with financial risk, profitability, tangibility and inflation.
Forte, Barros, and Nakamura (2013) investigated capital structure of Brazilian firms and found that profitability is significantly and negatively related to capital structure. Also, growth is positively and significantly related to leverage. Further, size is positively related and age is negatively related to the leverage.
Mac an Bhaird and Lucey (2010) explored 299 Irish enterprises and found that age, intangibility, collaterals and size are the significant determinants of capital structure.
They further found that ownership structure, size, age and collateral are similar across industry. Kouki and Said (2011) conducted research on 244 French listed companies and found that trade off theory, pecking order theory and market timings are not significant. Huang (2006) revealed an inverse relationship between leverage and profitability, non debt tax shield, growth opportunities and managerial shareholdings.

Determinants of Capital Structure -Pakistan
Khan, Jan, and Khan (2015)  found that profitability, liquidity, earnings volatility, and tangibility are negatively associated to leverage, while firm size is positively associated to leverage. Non-debt tax shields and growth opportunities do not appear to be significantly related to leverage.

Ahmad and Zaman (2013) analyzed textile sector listed on Karachi stock exchange of
Pakistan and revealed that profitability and size is significantly and negatively related to leverage while tangibility and growth were positively related to leverage. (2011)

 Agency Cost
Agency cost is one of the most important determinants of capital structure. It begins with the conflicts of interest between debt holders and equity holders (Myers, 1977). Firm having high agency cost have high cost of debt and thus leads to have lower debt ratio (Fama, 1980;Jensen, 1986;Titman, 1984).
Ho1: There is no significant relationship between agency cost and leverage of Pakistani firms.

Bankruptcy Cost
Bankruptcy cost depends on costs like legal fees, loss of sales, employees and suppliers and the probability of its happenings. If financing through debt increase, the probability of bankruptcy also increases and as a result bankruptcy cost increase. Firms with higher bankruptcy cost have lower debt.
Ho2: There is no significant relationship between bankruptcy cost and leverage of Pakistani firms.

 Non Debt Tax Shield
It is usually argued that company with more non debt tax shields should have less debt since the tax advantage of debt are comparatively less important (Akhtar & Oliver, 2009).
Ho3: There is no significant relationship between non debt tax shield and leverage of Pakistani firms.
 Profitability Myers (1984) argued that if a firm is more profitable then it will have more internal financing than external sources according to pecking order theory of capital structure. Therefore it can be proposed that the firms with higher profit have higher internal finance and hold less debt. Internal finance is less costly and easier whereas external finance is more costly. Hence it can hypothesize that there is an inverse relationship between profitability and leverage.
Ho4: There is no significant relationship between profitability and leverage of Pakistani firms.

 Size
Bigger firms usually have larger exposure to the public than smaller firms and consequently need to provide more information to consumers, creditors, suppliers, forecasters and government personals (Cooke, 1991). Larger firms have more resources to provide relevant information to stakeholders and as a result, larger firms have more debt with more attractive terms as compare to smaller firms. Hence, a direct association is expected between firm size and leverage. Empirical studies suggested size as a determinant of capital structure (Ferri & Jones, 1979;Scott Jr & Martin, 1975) and (Aggarwal, 1990).
Ho5: There is no significant relationship between size and leverage of Pakistani firms.

 Collateral Value Of Assets
Rajan and Zingales (1995) found that tangibility of assets or collateral value of assets is a determinant of capital structure. Corporations with more tangible assets are expected to have more debt because having more tangible assets gets debt easily on more favorable terms. On other side, Graham Jr. (1988) suggested that the corporations having high intangible assets have lower cost of borrowings cause better security for debt holders.
Ho6: There is no significant relationship between collateral value to assets and leverage of Pakistani firms.

 Business Risk
Business risk can be defined as the risk related with the future operations of the company. Firms with less business risk, (the risk that is connected with the upcoming business operations) are assumed more financial risks.
Ho7: There is no significant relationship between business risk and leverage of Pakistani firms.

 Growth
Theoretically it is suggested that the firm with higher growth rate will have lower debt in their capital structure. A company that grows fast invests huge amount in research and development.
Ho8: There is no significant relationship between growth and leverage of Pakistani firms.

 Age
As the firm grows, more information is available for the firm's probable viability in the future. More information cause less leverage in the capital structure of the firm.
Ho9: There is no significant relationship between age and leverage of Pakistani firms.

 Free Cash Flows
Jensen (1986) define free cash flows as the cash flow that is left after all positive NPV projects are funded. Harris and Raviv (1991) argued that firm with greater free cash flows will have lesser debt and vice versa.
Ho10: There is no significant relationship between free cash flows and leverage of Pakistani firms.  (Jensen, 1986), (Akhtar, 2004)

Business Risk
Volatility of Net Operating Income (Burgman, 1996) (Lee & Kwok, 1988), (Chaplinsky, 1984) Table 1 shows the proxies for dependent variable leverage and independent variables agency cost, free cash flows, growth, age, non debt tax shield, size, collateral value of asset, profitability, business risk, bankruptcy cost and foreign exchange risk.

Population
The initial data is collected from the publication of State Bank of Pakistan titled as "Financial Statement Analysis of the Companies (Non-Financial) listed at Karachi Stock Exchange". The publication includes only non financial firms. Some data is also collected from companies' websites and annual reports.

Sample
In this study, random stratified sampling technique is used and data is collected  parts, refined petroleum products, paper and board products, etc.  Table 3 represents the distribution of firm's years with respect to economic groups. Hence textile is the biggest economic group of Pakistan with total 624 firms' years with 38% in total. The second largest economic group is sugar and other food products which have 216 firms' years in total. Chemicals and pharmaceuticals and electric machinery and other manufacturing are also big economic groups with 176 and 160 firm's years respectively. In contrast refined petroleum products and paper, and paper board products are the smallest economic groups with 40 firms' years each.  Average bankruptcy cost is 9.3321.

Ethical Issues
As data is the secondary data, publically available on websites, no confidentiality or anonymity issues will arise.

Data Analysis
Initial data is collected and entered into Microsoft excel worksheet. The collected data has been entered accurately and systematically. The data for this research is the panel data means the data is the combination of time series data and cross sectional data therefore it has been organized accordingly and panel was created. In order to obtain the accurate empirical results, this study is using STATA 12 which is a very useful statistical tool for panel data. Different test have been applied in this research like descriptive statistics for the comparison of mean of variables and regression analysis for the relationship of variables.

Results And Finding
Sample for this study contains data across firms and over time so panel data analysis is appropriate. Panel data analysis has many advantages like it provide a hefty number of data points, increasing the degree of freedom, also decreasing the co-linearity among variables and helps in developing well-organized economic estimate (Hsiao, 1986). Further, panel data has advantage of make out and determine those effects that are simple not deductable in exclusive cross sectional or exclusive time series data (Baltagi, 1995). Moreover, Hsiao (1986) mentioned that panel data allows the application of variable intercepts models that initiate firm/industry type and/or time specific effects into the regression equation that minimize or evade the omitted variable bias. The most popular tools for analysis of panel data are fixed effect model and the random effect model. In this thesis the author is using the following decision making criteria for selection of the model presented by (Dougherty, 2011).
According to Figure 1, first of all, there is a need to perform both the fixed effect regression and the random effect regression if the data is selected randomly. Then    0.5495 0.0990 5.5500 0.0000 R-square within 0.0711, between = 0.3148, and overall = 0.1920 Wald Chi2 = 210.19, and Prob > Chi2 = 0.000, Variable is significant at * 1, ** 5, and ***10% level of significance (two-tailed) 360 0.000 R-square within 0.0878, between = 0.1398, and overall = 0.098 F Statistics = 13.27, and Prob > F = 0.000, Variable is significant at * 1, ** 5, and ***10% level of significance (two-tailed) Figure 1. Therefore, first of all these tests are applied to the sample of firms in Pakistan.
Both models are overall statistically good fit model as F test is significant in fixed effect model and Chi 2 is significant in random effect model in  Myers (1977). Size is a significant determinant of capital structure for all Pakistani firms (p-value = 0.010). Collateral value of assets is a significant determinant of capital structure (p-value = 0.000) and shows a positive relationship (coefficient = 0.023) with leverage indicating that if collateral value of asset increases, the leverage of the company also increases. Growth is a significant determinant of capital structure for the sample of Pakistani firms (p-value = 0.000) at the significance level of 1%. The results show a negative relationship with the leverage for firms (coefficient = -0.002). Age is significant determinant of capital structure for firms (p-value = 0.0000) and shows a negative relationship with the leverage. Business risk is not significant (p-value = 0.705).
Free cash flows is a significant determinant of capital structure for all firms (p-value = 0.0000) and for DCs (p-value = 0.0010). Free cash flows shows no relationship with leverage for (p-value = 0.705).

Pearson coefficient correlation
Correlations among variables can cause multicolinearity which may create problems in regression analysis. Table 6 shows correlations above 0.6 which explains that there is no multicolinearity among variables in however a modest correlation between free cash flows and non debt tax shield (0.702). Therefore, a multicolinearity test is further required to check any dependence among these variables in case of MNCs.

Multicollinerity test
When correlation among variables is high i.e. more than 0.6, either positive or negative, then the problem of multicolinearity may arise. To check whether multicolinearity exists among variables, a variance inflation factor (VIF) test is applied and hence the results show that ( Table 7) that there is no multicolinearity exists among variables.

Conclusion
This study used 202 Pakistani companies and analyzed eight years data for the annual periods of 2006 to 2013 and investigated the leverage composition of firms and the determinants of capital structure namely agency cost, bankruptcy cost, profitability, age, growth, collateral value of assets, non debt tax shield, free cash flows, size, business risk and foreign exchange risk. Fixed effect model was used to regress the variables. This study found that agency cost, growth, age, and size are significantly and negatively associated with the capital structure of Pakistan firms, however, collateral value of asset is significantly but positively associated with the capital structure of the firm. On the other hand, free cash flows, non debt tax shield, profitability, business risk and bankruptcy cost are not significantly associated with capital structure of the firms and are against the signaling theory and peaking order theory.
From this study, one can conclude that firms in Pakistan are using collateral values for getting more leverage out of their assets. Further, as agency cost, growth, age and size of the firm increases, shareholders prefer to invest rather than taking external debt.
Therefore, this study demonstrates imperative policy implications for managers and investors of the firms. This fact must be taken into account while making the finance mix decisions for securing the benefits of stakeholders.
Further research is required on several other factors that affect the capital structure of the companies like diversification, i.e. product diversification and geographical diversification. Both kinds of diversification not only affect capital structure of the companies but also have an impact on the profitability of the firms. Political risk is also a very important determinant of capital structure. Further, human capital is also a very vital determinant of capital structure in modern world. Very less work has been done on human capital. Firm's specific factors may also affect the leverage of the firm therefore