Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 Determinants of Capital Structure: Some Australian Service Sector Evidence Rafiuddin Ahmed1 Capital structure is one of the widely researched finance topics which generated a healthy body of literature in different countries. The key aim of this paper is to examine the determinants of leverage of seven service sector listed companies in the Australian stock exchange during the years 20122014. The study uses 57 individual firms and a total of 171 firm-year observations. The analyses of pooled and panel regression models reveal significant relationship between firm-specific and industry-specific determinants (factors) affecting leverage decisions of these firms. The factors are, however, dissimilar to factors observed in country-specific studies undertaken elsewhere. The reasons for differences in observed and extant factors in capital structure studies are explained, and the implications for research are outlined in this paper. Keywords: Capital structure, Australian service sector, panel data analysis, industry effect, firm characteristics. 1. Introduction Capital structure has been extensively researched since the seminal work of Modigliani and Miller (1958) was published. Since then, capital structure research has moved to examine firm value, managerial actions, and firm strategies to maximize income, shareholder value and growth of firms. Theoretical refinements to MM’s work (1958) are later clustered into different theories such as Trade-Off Theory, Pecking Order Theory, and Agency Theory (Jensen and Meckling 1976). The general results of these theories suggest an optimal capital structure has less than 100% leverage due to bankruptcy risks and optimal tax benefit threshold. The theoretical developments raised questions if costs and benefits are economically significant enough to have an appreciable impact on capital structure (Michaelas et al. 1999). This question is later empirically tested by relating firm characteristics such as size, profitability, growth rate, firm risk, and industry characteristics. The empirical studies range from robust cross country comparisons of capital structure to single country-specific studies. Multiple country studies are, however, not very recent. Fan et al (2012) examines capital structure differences of countries and observed intuitional and bond markets. Song et al (2004) examines the impact of country-specific factors on capital structure of 30 countries. John (2008) examines capital structures of 42 countries from all continents and finds different country-specific statistics of capital structures. However, Gitman’s (2004) study examines two clusters of seven countries each and finds a separate set of statistics for both sets. In single country-specific studies, researchers have used listed and non-listed companies or industry sectors and examined capital structures of firms in cross sections of an economy. For example, Abor (2007) uses firm characteristics of listed firms in Ghana, Shah and Khan 1 College of Business, Law and Governance, James Cook University, Townsville Campus, Australia. Email: rafiuddin.ahmed@jcu.edu.au 1 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 (2007) use firm characteristics of Pakistani firms, and finally, De Vries and Erasmus (2010) use firm characteristics of listed and unlisted firms in South Africa to explore the determinants of capital structure. There are many more studies in other country contexts but only some recent ones are listed here. Compared to studies in other countries, Australian studies are limited and very old. The Australian studies are confined to studying the imputation tax system and its effect of capital structure choice, firm value and shareholder values. Compared to the studies outside Australia, capital structure studies in Australia are still at a nascent stage and less conclusive (Qiu and La 2010). Australian capital structure studies are different and interesting Qiu and La (2010) due to the introduction of double taxation in 1987. Allen’s (1991) study on Australian firms’ capital structure reports financial flexibility and tax, Allen’s (1993) study examines theoretical assumptions of capital structure theories in the context of Australian firms (see also, Twite 2001). The study by Cassar and Holmes (2003) examines characteristics of Australian firms on capital structure choices (see also, Deesomsak et al. 2004). Finally, the most recent study by Qui and Ba (2010) reports the effect of firm characteristics of Australian firms over an extended period of time, over a 15year period (1992-2006). The Australian studies are too old, and none of the studies has examined the firm characteristics of the service sector and industry effects on capital structure. This current study is aimed to fill in this gap in the literature. The main contributions of this study to the literature on capital structure are the determination of firm characteristics of the service sector listed firms in the Australian Stock Exchange (ASX) and industry effects of three service sectors on leverage decisions. The findings are similar to the characteristics observed in earlier studies in other countries and the industry sectors are different from other studies. Recent evidence of service’s sector effects on capital in the Australian context is non-existent; this study is an addition to the global research on capital structure. The rest of the paper is organized as follows. In the next section, the theory and evidence of capital structure are presented. This follows the three consecutive sections on data, results and discussions. The final section rounds of the paper with conclusions and directions for further research. 2. Theoretical developments in capital structure research Since the development of MM theory in 1958 (Modigliani and Miller 1958) that capital structure has no relevance to firm valuation, other theories on capital structure and firm value have emerged. In 1963, MM (Modigliani and Miller 1963) refined their 1958 theory by adding taxes and relating it to firm value by stating that debt in a capital structure reduces a firm’s tax payable amount, improves income after taxes, resulting in a higher value of a firm than a firm without any debt in the capital structure. This development has later paved the way for the development of the trade-off theory which states that a firm’s optimal leverage is achieved by minimizing taxes, cost of financial distress and agency costs. Baxter (1967) argues that increased debt levels increase the chances of bankruptcy, and increase interest payable to the debtholders. A firm’s optimal leverage is where tax advantage from debt exactly equals the cost of debt. Kraus and Litzenberger (1973) argue that a firm’s market value declines if debt obligations are greater than its earnings. Deangelo and Masulis (1980) propose the static trade off theory and include other tax-minimizing offsets such as depreciation and investment 2 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 tax credits. They argue that firms weigh between tax advantages of debt with business risk (a cost). Their theoretical model proposes that a firm’s optimum debt level is where the present value of tax savings from debt equals the present value of costs of distress. Myer’s (1984) theoretical explanation of an asymmetric information hypothesis proposes different information held by firms’ internal and external stakeholders about firms’ income distribution plans (Ross, 1977). Thus, firms’ leverage level movements signal firms’ confidence levels, suggesting lower leverage as a poor signal about income and its distribution potential and vice versa. Pettit and Singer (1985) discuss the problems of asymmetric information and possible agency costs affecting firms’ demand and supply of credit. They argue that small firms possess higher level of asymmetric information due to financial constraints for sufficient disclosure of financial information to outsiders. This theory has laid the foundation for Pecking Order Theory (PET). Donaldson (1984) introduced the concepts and ideas of Pecking Order Theory (POT) which was later refined by Meyers (1984) and Myers and Majluf (1984). The fundamental premise of this theory is firm’s preferences for funding is stacked by a pecking order of risk preferences and corresponding costs. Thus, firms use the cheapest source of internal funds such as retained earnings, debt, convertible debt and preference shares) and the external equity (Myers 1984). The cost of sourcing extra funding is dependent on the extent of information asymmetries of risk perceptions emanating from differential information needs held by inside management and potential investors. In addition to a firm’s desire to source the cheapest fund to finance its needs, other factors such as the stage development of a firm such as a startup, a mature firm etc. influence the supply of funds (Mac an Bhaird and Lucey 2010). Agency theory (Jensen and Meckling 1976) addresses the fundamental problem of managing a firm’s capital structure from the cheapest source of funds. While common equity is an expensive source of fund, its use results in suboptimal firm value when equity holders insist on risk reduction from lower leverage usage. If managers’ and shareholders’ interests are not aligned, it is highly unlikely that optimal firm value is ever going to eventuate from managerial actions. The debt holders risk perceptions encourage them to ask for debt covenants or other costly debt shielding instruments. The tensions between the two subgroups of owners impose increased risk of monitoring by management resulting in costly monitoring and hence agency costs. A number of remedial measures such as reduction in consumption of resources when debt and hence, bankruptcy risks increase, increasing the stake of managers in a firm or increasing the leverage (Jensen 1986), commonly packed as ‘free cash flow hypothesis’. Free cash flow hypothesis proposes the adoption of measures to reduce free cash flow at managers’ disposal by increasing leverage (Stulz 1990) so that less cash flow is available for desired investment choices. 3. Empirical literature Empirical literature on capital structure covers robust cross country comparisons to single country-specific studies. First, some important cross country comparison studies are reviewed. Fan et al. (2012) compare the capital structure of 39 countries and find that degree of development of banking sector, equity and bond markets explained the heterogeneity of capital structures among the countries studied (see also, Hall et al. 2004). Song et al.’s (2004) study 30 OECD countries, however, find no support for country-specific variations as important determinants of heterogeneity of capital structure. Giannetti (2003) contends that 3 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 variations in findings of cross country comparisons of capital structure may be due to inappropriate sample selection such as consideration of only large firms in samples. John et al. (2008) criticize these studies for common country-level and firm level slopes and study for country-specific and firm level unique slopes and co-efficients. Their study uses 42 countries, from all continents and firms of all sizes. They find different country-specific slopes and different firm level slope and coefficients in these countries. In an earlier study, Gitman (2004) reports different slopes and co-efficient for seven firms each from developed and developing countries. While the studies reviewed above are aggregated country level studies, single country studies are also important in understanding firms operating in a unique regulator, economic and instructional frames. A large body of research examines single country-specific firm level factors affecting capital structure choices. Drobetz, Penza and Wanseried (2006) find that profitability, investment opportunities, tangibility of assets and volatility are important firm level determinants of capital structure (see also, Rajan and Zingales 1995). Vries and Erasmus (2010) examined firm and economic characteristics of listed South African firms affecting capital structures in the industrial sector. The study uses panel data over a period of 14 years (1995-2008) and 2684 firm year observations (170 listed and 110 unlisted firms). The results suggest that asset structure (Fixed assets/Total assets) and natural log of sales [ln(sales)] are two most important determinants of capital structure of South African listed and unlisted industrial firms. The study also reports different slopes and coefficient values by subsets of listed and unlisted firms, suggesting the differences due to ‘survivorship bias’ (De Vries and Erasmus 2010, p 6). Australian capital structure studies are different and interesting (Qiu and La 2010) due to the introduction of double taxation in 1987. Allen’s (1991) study on Australian firms’ capital structure reports financial flexibility and tax as two main concerns of capital structure choices of Australian firms. Allen (1993) finds profitable Australian firms’ inclination to use less debt, confirming signaling and pecking order theory and contradicting trade-off theory. Twite (2001) reports significant debt ratio reductions, more reliance of capital sourcing following the tenets of pecking order theory and signaling theory but not so much about trade-off theory. The study by Cassar and Holmes (2003) reveals fast-growing and larger firms use more debt but firms with less tangible assets use less debt. Their survey fails to find any significant effect on a firm’s capital structure (debt ratio). Deesomask et al’s (2004) study fails to find profitability, growth, and risk as explanatory factors of capital structure differences in Australia. Qiu and La (2010) reports in their study of Australian firms over a 15 year period (1992-2006) that size is not statistically significant but tangibility of assets and profitability have positive relation and growth prospects and business risks have negative relationship with leverage. The Australian studies are too old and none of the studies have examined the firm characteristics of the service sector and industry effects on capital structure. This current study is aimed to fill in this gap in the literature. 4. Data and procedures This study is based on data collected from Datanalysis database. The database provides financial information about Australian companies operating in different sectors. The current study uses only three years of complete data (2012-2014) on 60 service companies that match the search criteria. Three of these companies are later excluded due to incomplete 4 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 information for analysis. Thus our final sample comprises 57 companies (panel data) and 171 firm year observations (pooled data). The companies comprise seven sectors of the Australian economy and have positive values for all search entries. Only listed firms from these seven sectors are included, so the sample may have survivorship bias (De Vries and Erasmus 2010). Michaelas et al. (1999) argue that surviving firms comprise a material component of an economy, so only listed firms will provide important insights to our analyses of capital structure information. Table below shows the simple descriptive statistics of the variables (n =171), disregarding the panel information. 5 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 Table 1: Descriptive statistics of dependent and independent variables Longterm liability Total liability Current liability Asset growth Operating revenue Profitability Asset tangibility Depreciation ratio Growth Mean 18.0949 18.9479 17.9582 0.2577 19.4497 0.9033 19.8949 0.0179 1.5834 Median 17.7616 19.1021 18.1484 0.1443 19.0705 0.9065 20.0000 0.0140 0.9334 Maxim um 22.4236 23.5444 23.4492 2.3135 23.9159 0.9752 24.1617 0.0893 25.000 Minim um 13.2375 13.5424 9.3955 0.0000 14.6831 0.7723 14.5960 0.0001 0.0000 Std. Dev. 2.1879 2.5513 3.2371 0.3161 2.1633 0.0309 2.3366 0.0144 2.6384 Probability 0.0351 0.0141 0.0428 0.0000 0.0694 0.0000 0.0311 0.0000 0.0000 4.1 Measurement of variables: Dependent variables There are different ways of calculating leverage in the capital structure. Total debt ratio: Total debt to total assets (Michaelas et al. 1999) but Delcoure (2007 , p 269) has modified it by suggesting the book value of debt to book value of debt plus the book value of equity. Long term debt ratio: long term debt by total assets (Michaelas et al. 1999, Delcoure 2007) Short term debt ratio: short term debt to total assets (Michaelas et al. 1999, Delcoure 2007) Independent variables Size: Natural logarithm of Sales (Titman and Wessels 1988, Michaelas et al. 1999, Delcoure 2007, De Vries and Erasmus 2010) Profitability: Earnings before interest and taxes divided by Total Assets (Michaelas et al. 1999, De Vries and Erasmus 2010) Growth: Percentage growth in total assets over a three-year period (Titman and Wessels 1988, Michaelas et al. 1999, Delcoure 2007) Depreciation ratio Depreciation expenses divided by total Assets (Michaelas et al. 1999, Delcoure 2007) Tangibility of assets: Total fixed assets divided by total assets (Michaelas et al. 1999, Delcoure 2007) Industry dummies: Selected sector dummy is set to ‘1’ value, otherwise a ‘0’ value for each subject. The process is repeated for six sectors and one sector is kept as a benchmark sector for comparison purposes (no dummies assigned (Delcoure 2007). 4.2 Model specification The study uses two different models, a constant co-efficient model (pooled regression model) and a panel regression model (time series regression). The pooled regression model ignores time and industry effects, assumes a constant value for the intercept for the entire sample, and the slope co-efficient are the same for the sample (171 observations in this study). The model is specified below. LGit = β0 + β Xit + ϵit 6 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 where LGit = the measure of leverage of a firm I at time t; β0 = the intercept of the equation; β = the change coefficient of Xit = the different independent variable for a leverage of a firm i at time t; i = the number of the firms (sample size) = 1, 2, …57, and t = the time periods (1, 2...N). In order to examine the relationship between leverage and firm level variables, panel data is used. Panel data uses time and cross-sectional data, increases data points in a single study, increases data points, increases degree of freedom which in turn reduces multi-collinearity among explanatory variables leading to improved econometric estimates (Shah and Khan 2007). In order to choose the most appropriate model that takes into account industry and time effects, Hausman Test is used for model selection. The equations of the models, the fixed effect and the random effects model, are as follows: Fixed effect model: LGit = aoi + β 1 X 1it + β2 X 2 it…+ U it The subscript i in the intercept represents different intercepts across sectors but the ‘i’ used with the slopes assumes the same slope across different industry sectors. Therefore, the following Random effects model is developed for this study. Random effect model: LGit = μ+ + β X it + (Ui + ϵit) Where LGit is the i th individual at time, X it is a vector of independent variable, β is a vector of parameters, and (Ui + ϵit) is the error term. 5. Results and discussions The table below shows the best pooled regression model (backward method) including the industry dummies. The best pooled model in Table 2 below shows only the significant determinants of capital structure (at 5% level of significance). Table 2: Pooled regression analysis Model (Constant) Operating_revenue TangibiltyofAssets Dep_ratio Asset_growh Materials softwareequop Telecom Co-efficientsa Unstandardized Standardized Coefficients Coefficients Std. B Beta t-value Error -1.663 0.294 -5.664 0.517 0.040 0.511 12.961 0.480 0.037 0.513 13.009 4.676 2.003 0.031 2.334 0.160 0.089 0.023 1.796 -0.148 0.070 -0.027 -2.117 0.186 0.082 0.031 2.263 0.235 0.111 0.028 2.115 Probability 0.0000*** 0.0000*** 0.0000*** 0.02100** 0.074000* 0.03600** 0.02500** 0.03600** 7 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 a. Dependent Variable: B_LONGTERMLIABILITY *** represents 1% level of significance, **represents 5% level of significance, * represents 10% level of statistical significance, and * represents significance at 10% level. The best model in Table 2 shows four firm level explanatory variables and three industry sector specific factors at different significance levels marks with the number of asterisk/s (*). A pooled regression model does not consider time effect, so two additional models are run in the Eviews software. The first model uses only firm-specific panel data and reports three statistically significant independent variables. Earnings Before Interest and Taxes (EBIT), Depreciation Ratio, and profitability, explain the variability of long term liability (in absolute terms) of sample. This model, however, has not considered time and industry effects at the same time, so panel models are used to determine time and industry effects. Table 3: Panel regression analysis results (without industry dummies) Variable t-statistic Probability Constant Coefficient -2.076712 -2.261851 0.0250 Depreciation ratio 4.893158 2.237506 0.0266** Size Tangibility 0.525015 0.457126 12.70309 12.28565 0.0000** 0.0000** Profitability 0.890949 0.901469 0.3687 Growth -0.016724 -1.589905 0.1138 ** represents 5% level of statistical significance Table 3 above shows that three independent variables, size, depreciation ratio (DEP_RATIO) and tangibility of assets, are important determinants of capital structure of Australian service sector firms. In the table, only long term liability is used as the best descriptor of leverage. The model above, however, does not consider industy and time effects. Table 4 below reports the output of a panel model with industy dummies. Based on Hausman test, the output of the random effects model is presented (null hypohesis that random model is appropriate is accepted at 5% level, p= 0.57). Table 4: Panel regression analysis results (with industry dummies) Variable Co-efficient t-statistic probability C 12.605 1.8614 0.0645 EBIT 0.9279 2.3206 0.0216** ASSET_GROWH 0.0671 0.8512 0.3959 TANGIBILTY_OFASSET... 0.2867 0.8019 0.4238 DEP_RATIO 15.6194 2.1387 0.0340** PROFITABILITY -17.8994 -2.3637 0.0193** PHARMA 0.2699 1.3153 0.1903 SOFTWAREEQUOP 0.3510 2.2022 0.0291** UTILITIES 0.3172 1.7542 0.0813* TELECOM 0.3329 1.6533 0.1002 8 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 ENERGY -0.0876 -0.5497 0.5833 HEALTHCAREEQUIPMEN... 0.1308 0.9270 0.3553 *represents 10% level of significance, ** represents 5% level of significance Table 4 above shows that three independent variables, EBIT, depreciation ratio (DEP_RATIO) and profitability (PROFITABILITY), are found as important determinants of capitl structure of Australian service sector firms. In the table, only long term liability is used as the best descriptor of leverage. The model above, considers industy effects but no time effect due to an insufficient number of samples. Out of seven industry sectors, only software sector (significant at 5% level) and utilities sector (significant at 10% level) affects capital structure signifcantly. The results of the tables above suggest EBIT, depreciation ratio and profitability as three generic significant determinant of capital structure of Australian service sector firms. The first factor, profitability, is related to a firm’s choice of sourcing capital during the times of need. Following Myers (1984) and his Pecking Order Theory (PET), a firm prefers internal to external financing when it comes to raising additional capital. Therefore, a negative correlation is expected between a firm’s leverage and profitability. In the first model above, profitability has a positive co-efficient, confirming the PET expectations. Whereas the tradeoff theory (Jensen 1986)(Jensen, 1986) suggests tax deductibility and tax bill reductions as the prime reasons for debt usage. Firms with high profitability are more affluent and hence a positive association between profitability and leverage is expected. The result confirms the trade-off theory hypotheses. The co-efficient of profitability is not statistically significant at any level of statistical significance as three standardized beta coefficients are negative. However, when sector dummies are used in the second panel model in Table 4, the results show a negative co-efficient for profitability in line with the predictions of trade-off theory of capital structure. The second important determinant is the depreciation ratio in Table 3 and 4. Depreciation and other non-debt tax deductible items are substitutes for tax shields on debts (DeAngelo and Masulis 1980) So a negative association between depreciation ratio and leverage is expected. The outputs of the two models in Table 3 and 4 reveal positive beta coefficients for the debt ratio (all significant at 5% level) suggesting that Australian firms in the sample are highly levered, that is, use more long term debts to fund their assets. The third determinant, tangibility of assets, is statistically significant (co-efficient value 0.4571 and p = 0.0000). The positive coefficient means for each percentage point increase in tangible assets, debt also increases by 0.4571%. This finding is consistent with the predictions of Jensen and Meckling’s (Jensen and Meckling 1976) agency theory and Myer’s (1984)tradeoff theory. The finding is also consistent with findings in other countries and different from the findings in other countries. The results suggest that Australian firms are still insufficiently levered. As a result, the owners are not earning enough from their investments, and the value of the firms in the service sector has not reached to the optimal level. A mirror image of this explanation is that Australian firms are using more equity to fund their operations, which means less debt is used, implying suboptimal earning from the existing configuration of capital structure. This can be improved by increased borrowing level. The fourth factor, growth, has a negative co-efficient (-0.016721) and is not statistically significant at 5% level (p = 0.1138) but insignificant in the second model in Table 3 and Table 4 (coefficient value is 0.2867, p= 0.4238). This result suggests that Australian service sector 9 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 firms are not funding their operations from long-term debt. The finding is dissimilar to Shah and Khan (2007). Some plausible explanation can be the developing nature of the sector and increased risk perceptions by debtors to lend money to the service-sector firms. Borrowing by rising firms may be perceived risky and may attract a high risk premium, and resultant high debt interest rate demands by creditors. The second explanation may be the risk perception of managers and reluctance to take risks when the service-sector firms are growing. This is to save managers’ own jobs as well and to invest in a stable project only ( see, for example, Shah and Khan 2007). Size is an important determinant of capital structure in almost all studies. In this study, size is measured by natural log of sales. The coefficient value is positive (0.5250) and statistically significant at 5% level (p= 0.0000). The results suggest that Australian service sector organizations are enjoying the benefits of increased revenue from leverage, that is, for each percentage point of an increase in revenue, long-term debt increases by 0.52%. Prior literature has two opposing expectations, a positive relationship (Titman and Wessels 1988) and a negative relationship (Rajan and Zingales 1995). A positive association is expected when investors have confidence in a firm and are willing to lend without the fear of bankruptcy, so debt is cheaper (Shah and Khan 2007) whereas, when a negative relation is expected, it implies less confidence and more asymmetric information held by firms or a rising sector (Shah and Khan 2007). The results support the view of Titman and Wessels (1988), suggesting that the Australian service sector is quite established, it has the confidence of the debt market, and that the firms in the sector are able to utilize debt effectively for expansion of their operations. Sector dummies are used to study industry effects on leverage decisions. Industries from two different sectors reveal a significant level of leverage in their capital structure. These two sectors, software and utilities, are approximately 35% more levered than the benchmark sector (the materials sector), suggesting the firms in this sector are more established and profitable. 6. Conclusions The current study examines firm characteristics and industry sector effects of the Australian service sector listed firms in the ASX. Using pooled and panel regression models, the study reports some significant firm characteristics and industry sectors that affect capital structure choices. Due to variations in the way capital structure is defined in extant research, the results of this study also report different determinants when a capital structure is defined differently, and control variables are changed to get the results. The pooled regression model reports some control variables, long term debt and some industry sectors as significant determinants of capital structure of the firms. The panel regression (without industry effect) reports tangibility, size and depreciation as important determinants of capital structure of the firms. When industry effect is considered, a different set of control variables is used to get a meaningful result. The random effect model suggests that Earnings before interest and taxes (EBIT), profitability and depreciation ratio are significant firm level determinants and Software and Utilities sectors are significantly more levered than the benchmark sector. The study is based on 57 firms, over two years, 2012 to 2014 and a total of 171 firm-year observations. A larger sample size, either by including more firms or more number of years, may reveal more robust results than reported in this paper. The study uses only listed 10 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 companies in the service sector, so it suffers from ‘survivorship bias’ (De Vries and Erasmus 2010). To overcome this limitation, inclusion of non-listed companies may be considered. Time dummies tell important information about the movements of leverage structure over time. Due to an insufficient number of observations, time dummies cannot be included in panel regression analysis. Inclusion of more time periods may overcome the time dummy related data processing problems. The exploratory nature of this study is based on only six independent variables. Data availability on more independent variables will obviously overcome this shortcoming. 7. REFERENCES: Abor, J., 2007. Corporate governance and financing decisions of Ghanaian listed firms. Corporate Governance: The international journal of business in society, 7, 83-92. Allen, D.E., 1991. The determinants of the capital structure of listed Australian companies: the financial manager's perspective. Australian Journal of Management, 16, 103-128. Allen, D.E., 1993. The pecking order hypothesis: Australian evidence. Applied Financial Economics, 3, 101-112. Baxter, N.D., 1967. Leverage, risk of ruin and the cost of capital*. the Journal of Finance, 22, 395-403. Cassar, G. and Holmes, S., 2003. Capital structure and financing of SMEs: Australian evidence. Accounting and Finance, 43, 123-147. De Vries, A. and Erasmus, P., 2010. Determinants of capital structure: A South African study. Corporate Ownership and Control, 8, 591-599. Deangelo, H. and Masulis, R.W., 1980. Optimal capital structure under corporate and personal taxation. Journal of financial economics, 8, 3-29. Deesomsak, R., Paudyal, K. and Pescetto, G., 2004. The determinants of capital structure: evidence from the Asia Pacific region. Journal of Multinational financial management, 14, 387-405. Delcoure, N., 2007. The determinants of capital structure in transitional economies. International Review of Economics & Finance, 16, 400-415. Donaldson, G., 1984. Managing corporate wealth. New York: Praeger. Drobetz, W. and Wanzenried, G., 2006. What determines the speed of adjustment to the target capital structure? Applied Financial Economics, 16, 941-958. Fan, J.P., Titman, S. and Twite, G., 2012. An international comparison of capital structure and debt maturity choices. Journal of Financial and quantitative Analysis, 47, 23-56. Giannetti, M., 2003. Do better institutions mitigate agency problems? Evidence from corporate finance choices. Journal of Financial and quantitative Analysis, 38, 185-212. Hall, G.C., Hutchinson, P.J. and Michaelas, N., 2004. Determinants of the capital structures of European SMEs. Journal of Business Finance and Accounting, 31, 711-728. Jensen, M.C., 1986. Agency cost of free cash flow, corporate finance, and takeovers. Corporate Finance, and Takeovers. American Economic Review, 76. Jensen, M.C. and Meckling, W.H., 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of financial economics, 3, 305-360. Kraus, A. and Litzenberger, R.H., 1973. A state‐preference model of optimal financial leverage. the Journal of Finance, 28, 911-922. Mac an Bhaird, C. and Lucey, B., 2010. Determinants of capital structure in Irish SMEs. Small business economics, 35, 357-375. 11 Proceedings of World Business, Finance and Management Conference 14 - 15 December 2015, Rendezvous Grand Hotel, Auckland, New Zealand ISBN: 978-1-922069-91-7 Michaelas, N., Chittenden, F. and Poutziouris, P., 1999. Policy implications arising from the impact of the last recession on the capital structure of UK SMEs. Environment and Planning C: Government and Policy, 17, 411-430. Modigliani, F. and Miller, M.H., 1958. The cost of capital, corporation finance and the theory of investment. The American economic review, 261-297. Modigliani, F. and Miller, M.H., 1963. Corporate income taxes and the cost of capital: a correction. The American economic review, 433-443. Myers, S.C., 1984. The capital structure puzzle. the Journal of Finance, 39, 574-592. Myers, S.C. and Majluf, N.S., 1984. Corporate financing and investment decisions when firms have information that investors do not have. Journal of financial economics, 13, 187221. Pettit, R.R. and Singer, R.F., 1985. Small business finance: a research agenda. Financial management, 47-60. Qiu, M. and La, B., 2010. Firm characteristics as determinants of capital structures in Australia. International Journal of the Economics of Business, 17, 277-287. Rajan, R.G. and Zingales, L., 1995. What do we know about capital structure? Some evidence from international data. the Journal of Finance, 50, 1421-1460. Shah, A. and Khan, S., 2007. Determinants of capital structure: Evidence from Pakistani panel data. International review of business research papers, 3, 265-282. Song, J. and Philippatos, G., 2004. Have we resolved some critical issues related to international capital structure? Empirical evidence from the 30 OECD countries. Stulz, R., 1990. Managerial discretion and optimal financing policies. Journal of financial economics, 26, 3-27. Titman, S. and Wessels, R., 1988. The determinants of capital structure choice. the Journal of Finance, 43, 1-19. Twite, G., 2001. Capital structure choices and taxes: evidence from the Australian dividend imputation tax system. International Review of Finance, 2, 217-234. 12