Capital Structure and Corporate Performance: An Empirical Analysis from Central Europe

Abstract

The capital structure and its indicators play a significant role in corporate finance. The aim is to estimate business performance using selected indicators describing primarily the capital structure, asset structure, or liquidity of transport companies in Central Europe. The total sample consists of almost 4000 small and medium-sized enterprises in the transport sector. This data is collected from Amadeus Bureau van Dijk. The results show that six out of ten variables are statistically significant predictors affecting business performance; two out of the six indicators are categorical variables, such as the company size classified into small and medium enterprises and the country divided into the Czech Republic, Hungary, Poland, or Slovakia. We find that Hungarian medium-sized enterprises show higher profitability than other enterprises, assuming other factors are unchanged. Finally, the results demonstrate that a high debt ratio and a high share of non-current assets in total assets have a negative impact on corporate performance in contrast to the current ratio and the share of cash and cash equivalents in total assets. In other words, liquidity and cash and its equivalents have a significant role in increasing business performance. These findings are specific because, generally, high liquidity does not positively impact performance.

1. Introduction

In the V4 countries, small and medium-sized companies in the transport sector provide a range of services, including freight and passenger transportation, logistics, warehousing, and distribution. These companies employ a significant portion of the workforce in the transport sector and contribute to the region’s overall economic growth and development. Small and medium-sized companies in the transport sector face various challenges, including competition from larger companies, increasing regulatory requirements, and the need to adopt new technologies and business models. However, these companies are also agile and adaptable, and they often have a strong focus on customer service and niche markets. Overall, small and medium-sized companies in the transport sector are essential drivers of economic growth and innovation in the V4 countries, and their continued success is essential for the continued development of the region’s economy. In the transport sector, where large investments in fixed assets such as vehicles, terminals, and other infrastructure are required, the choice of capital structure can significantly affect the cost of capital, risk, and profitability of a company.

A company’s capital structure refers to the mix of debt and equity used to finance its operations and growth. The trade-off theory is a financial concept that explains the relationship between a company’s debt and equity financing. It suggests that there is an optimal level of debt that a company should maintain and that too much or too little debt can be detrimental to the firm’s overall value. On the other hand, the pecking order theory of capital structure suggests that companies follow a specific order in their financing decisions, with internal funds (retained earnings) being the preferred source, followed by debt and equity. The theory states that companies prefer debt financing when they have positive cash flows and equity financing when they have negative cash flows. Both trade-off theory and pecking order theory are widely used to explain the capital structure decisions of firms and help financial managers determine the optimal mix of debt and equity to use in financing their operations. However, in practice, firms may deviate from these theories due to various factors such as market conditions, regulatory restrictions, and management’s risk tolerance. Brusov and Filatova [1] analyzed all existing theories of capital structure with an emphasis on advantages and disadvantages. The most relevant papers about capital structure and corporate performance are from Berger and di Patti (2006) [2], Margaritic and Psillaki (2010) [3], Simerly and Li (2000) [4], King and Santor (2008) [5], Chaganti and Damanpour (1991) [6], Nadiri and Mamuneas (1994) [7], Ofek (1993) [8], Andrews (2010) [9], Ebaid (2009) [10], and Gleason, Mathur and Mathur (2000) [11] according to the total number of citations in Web of Science.

This paper aims to create a model for predicting business performance using indicators describing capital indicators together with control ratios from previous research. This research deals with different industries. The purpose of this paper is to provide new insights into capital structure, asset structure, and firm size on corporate performance for transport companies using a universal model for Visegrad Group countries. Moreover, this paper identifies potentially statistically significant indicators for modeling business performance. These findings will point to a preferred capital structure policy among indicators describing the capital structure, method of financing, and corporate performance.

We find that the modeling of corporate performance based on capital structure and control indicators is not typical for the transport sector. This scientific gap motivates the creation of a specific universal regression model considering primarily the impact of capital structure and their control variables on business performance in transport companies in V4; this model will help all domestic or foreign investors.

The added value of this paper lies in the design of a regression model for small and medium-sized enterprises in the transport sector in Central Europe so that managers, owners, investors and banking institutions do not have to rely on models from earlier research based on smaller samples and other industries in different countries. This model estimates corporate performance using a narrow range of indicators describing capital structure along with other indicators commonly used with capital structure indicators. Our research is based on a wide range of theoretical and empirical knowledge from previous research, especially over the last decade.

This literature review concentrates on theoretical and empirical knowledge about capital structure and its potential determinants for modeling business performance from previous research. This section summarizes the financial indicators. We find that six main indicators of capital structure, such as debt ratio, long-term debt ratio, current ratio, short-term debt ratio, the share of non-current assets to total assets, and company size as the natural logarithm of total assets, are often used to model performance. Moreover, we also use other indicators such as country, company size divided into small and medium-sized enterprises, and legal form divided into private or public enterprises. The methodology presents a sample of almost 4000 enterprises from the Central European region, including the Czech Republic, Poland, Hungary, and Slovakia. The dataset includes several financial indicators with a formula, acronym, and reference to previous research. In this section, we explain the individual steps for creating a relevant regression model for business performance modeling. The results present a model for all countries. Finally, we compare our theoretical and empirical findings with research on capital structure from previous research, especially from the European region.

2. Literature Review

Capital structure and its influence on value creation have been decisive theses since the semester work by Modigliani and Miller (1958) [12]. Current research deals with capital structure during the COVID-19 pandemic [13,14,15,16,17,18,19]. The theory of capital structure is divided into trade-off theory, pecking order theory, and market timing theory. Firstly, the trade-off theory is based on research on taxes, corporate bankruptcy, and the cost of financial distress to determine the optimal capital structure broken down into own and external resources. Serrasqueiro and Caetano (2009) [20] added that businesses should achieve debt levels that maximize the benefits of tax shields and minimize the risk of bankruptcy. Secondly, Culata and Gunarsih [21] summarized the findings of the pecking order theory, according to which companies prefer retained earnings for financing. In general, earnings, short-term securities, debt, and preference shares are more favored than ordinary shares. This theory uses debt resources under the pressure of insufficient internal resources. Thirdly, Jahanzeb et al. (2013) [22] and Miglo (2010) [23] explained that financial managers make their decisions on issued shares depending on market performance.

Overall, the theories of capital structure are important frameworks that attempt to explain how firms decide to finance their investments and operations. Two popular theories are the trade-off theory and the pecking order theory. However, these theories have some shortcomings that limit their usefulness in predicting and explaining firm behavior. Firstly, the trade-off theory suggests that firms choose their optimal capital structure by balancing the benefits of debt (such as tax shields and lower costs of capital) against the costs of debt (such as financial distress and agency costs). However, this theory assumes that firms have perfect information and can make optimal decisions about their financing needs. In reality, firms may face uncertainty and imperfect information, which can make it difficult to accurately assess the costs and benefits of debt. Additionally, the trade-off theory assumes that firms can easily access the capital markets to raise debt and equity capital, but in practice, market conditions and institutional factors may limit firms’ ability to access financing. Secondly, the pecking order theory suggests that firms prefer to finance their investments using internal funds, followed by debt and then equity. This theory assumes that firms have asymmetric information about their true value and that external financing signals a negative message to the market. However, empirical evidence suggests that firms do not always follow the pecking order and may issue equity even when they have positive cash flows. Additionally, the pecking order theory does not explain why firms might choose to issue debt rather than equity in certain circumstances.

García-Gómez et al. (2021) [24] argue that current theoretical and empirical knowledge of capital structure is still inconclusive. Their research is focused on companies from air services, hotels and resorts, leisure and recreation, and tourist services based on data from the Thomson Eikon database. These authors suggest that companies implement alternative business strategies for excessive leverage. This effect negatively affects business performance. Nguyen and Nguyen (2020) [25] examined the impact of capital structure on the performance of state-owned and non-state-listed companies on the Vietnam Stock Exchange. Their research focuses on examining the relationship between the share of long-term and short-term debt, respectively, on the total assets and performance of Vietnamese companies. Silva Serrasqueiro and Rêgo Rogão (2009) [20] drew similar data on Portuguese stock exchanges from the Documentation Center of Euronext Lisbon and the Finbolsa database. The results show that capital structure has a statistically significant negative impact on corporate performance. In addition, this impact is stronger in state-owned than non-state-owned enterprises. These empirical findings provide a new managerial view of state and non-state enterprises to improve business performance. Later, Mursalim et al. (2017) [26] identified which determinants affect the corporate performance of Asian companies in Indonesia, Malaysia, and Thailand. The overall sample consists of 94 Indonesian, 153 Malaysian, and 74 Thai companies from 2008 to 2012. Research has shown that company size and volatility play a dominant role in explaining variations in capital structure. In addition, they demonstrate that capital structure is statistically significantly related to corporate performance. Khémiri and Noubbigh (2018) [27] found that the Generalized Method of Moments (GMM) and the quadratic method support the trade-off and pecking order theories. The results show that there is a significant inverse relationship between debt ratio and performance; key factors include macroeconomic factors and leverage. Furthermore, Vijavakumaran and Vijavakumaran (2019) [28] explained the impact of corporate governance on the capital structure of Chinese companies using GMMs. Spitsin et al. (2021) [29] drew data on financial indicators from high-tech industries and services based on the international classification NACE from Spark Information Systems from 2013 to 2017. The total sample consists of 1826 enterprises broken down into 585 enterprises in high-tech industries and 1241 companies in high-tech services. Research aimed to determine the optimal capital structure using a set of variables such as the company size based on the natural logarithm of sales adjusted for inflation, the share of tangible assets in total assets, and current liquidity, with the independent variable being the debt ratio. The results show that the debt ratio of Russian high-tech companies exceeds the optimal level. Sudiyatno et al. (2020) [30] argued that capital structure and ownership directly affect business value. These authors found that management decisions and company size positively impact profitability as opposed to capital structure. On the other hand, capital structure and ownership have a negative impact on the value of the company as opposed to the company size. Nenu et al. (2018) [31] summarized theoretical and empirical findings on the impact of capital structure on risk and corporate performance. Their results show that the debt ratio is positively correlated with corporate size and stock price volatility. On the other hand, the debt structure has a different impact on the performance and development of market share prices.

Silva Serrasqueiro and Rêgo Rogão (2009) [20] explained that companies should achieve a level of indebtedness that maximizes the benefits of tax shields and minimizes potential bankruptcy based on the trade-off theory. These authors analyzed decisions on the capital structure of small and medium-sized enterprises to meet the relevant assumptions of the selected capital structure theory, such as the trade-off theory and pecking order theory. Silva Serrasqueiro and Rêgo Rogão (2009) [20] found a negative relationship between debt ratio and corporate performance, based on which SMEs prefer internal rather than external sources. In addition, the results show a statistically significant negative relationship between the debt ratio and the age of the company. In other words, a company with a long history does not tend to use foreign sources. On the other hand, there is a statistically significant relationship between the debt ratio and the company size, which indicates a diversification of activities to minimize the likelihood of the company going bankrupt. In conclusion, Silva Serrasqueiro and Rêgo Rogão (2009) [20] summarized that small and medium-sized enterprises prefer the pecking order theory for the negative relationships between the debt ratio and the performance or size of the enterprise. These findings are typical of the use of internal financing, especially for start-ups. In addition, the positive relationship between the debt ratio and the company size is based on easier access for larger enterprises to foreign resources on more favorable terms. Fernandes et al. (2018) [32] found that medium-sized enterprises have a higher average efficiency than small enterprises. The results show that efficiency increases with the share of current assets in current liabilities. In addition, more efficient businesses prefer higher leverage because higher efficiency will reduce the cost of potential bankruptcy. In other words, leverage positively impacts business efficiency according to the pecking order theory. These authors find that short-term debt is negatively associated with the size of the business, reflecting the high transaction costs in small businesses.

Vătavu (2015) [33] analyzed the relationship between the capital structure and the performance of Romanian companies. This research shows that performance is higher in the use of equity as opposed to debt financing. However, manufacturing companies do not have sufficient internal resources to make investments. In a period of increased taxes and inflation, profitable companies are getting rid of assets, which contributes to reducing costs.

Research on capital structure in Central Europe is negligible. Grabinska et al. (2021a) [34] and Grabinska et al. (2021b) [35] developed research on capital structure in specific sectors of the Polish economy. Valaskova et al. (2021) [36] and Skerlikova and Rudolfova (2015) [37] examined external and internal factors as basic elements in deciding on the capital structure of Slovak and Czech companies. On the other hand, Jindrichovska et al. (2013) [38] examined the influence of the independent variables such as debt ratio for the previous year, growth opportunities, asset structure, profitability, company age, and liquidity on the capital structure of 260 small and medium-sized enterprises from the Czech Republic for 2004–2011. These authors [39] extended their research to include working capital management and its impact on corporate performance as one of the pillars of corporate finance. Finally, Wieczorek-Kosmala et al. (2021) [40] analyzed the factors that determine the performance of unlisted energy companies from Hungary, Poland, Slovakia, and the Czech Republic. This research supports the assumptions in the pecking-order theory. On the other hand, it reveals the assumptions of the trade-off theory for short-term debt using a regression analysis based on corporate data from 2015–2019.

The capital structure and its impact on business performance are determined together with other control variables based on broad theoretical knowledge from several studies by García-Gómez et al. (2021) [24], Ngatno et al. (2021) [41], Spitsin et al. (2021) [29], Nenu et al. (2018) [31], Vătavu (2015) [33], Tousek et al. (2021) [42], Vuković et al. (2022) [43], Lehenchuk et al. (2022) [44], Wieczorek-Kosmala et al. (2021) [40] and others. We find that the capital structure ratios are often expressed as total liabilities to total assets, long-term liabilities to total assets, and current liabilities to total assets. On the other hand, control indicators include current assets to current liabilities, non-current assets to total assets, the natural logarithm of total assets, and cash and cash equivalent to total assets. We find that the total debt ratio has a negative impact on business performance according to Ngatno et al. (2021) [41], Spitsin et al. (2021) [29], Vătavu (2015) [33], Vuković et al. (2022) [43], and Wieczorek-Kosmala et al. (2021) [40]. This impact is evident in several sectors, such as the manufacturing sector, hi-tech sector, advertising sector, and energy sector. Long-term debt ratio has a negative impact on performance, according to Tousek et al. (2021) [42] and Vuković et al. (2022) [43]. However, Nenu et al. (2018) [31] explain that this indicator is not statistically significant according to their fixed effect model. Finally, the short-term debt ratio has a significant negative effect, according to Vuković et al. (2022) [43]. Capital structure together with other control indicators from liquidity, asset structure, and company size, are often used to model business performance in various industries. Firstly, Vătavu (2015) [33], Tousek et al. (2021) [42], and Wieczorek-Kosmala et al. (2021) [40] demonstrated that current ratio has a negative effect on corporate performance compared to Vuković et al. (2022) [43], Tousek et al. (2022) [42], and Wieczorek-Kosmala et al. (2021) [40]. However, Spitsin et al. (2021) [29] found that this indicator is not statistically significant. Secondly, non-current assets to total assets also have a negative impact on corporate performance, according to Spitsin et al. (2021) [29], Vătavu (2015) [33], and Vuković et al. (2022) [43]. Thirdly, corporate size according to the natural logarithm of total assets and its impact on corporate performance is not unambiguous because García-Gómez et al. (2021) [24], Vuković et al. (2022) [43], Lehenchuk et al. (2022) [44], and Wieczorek-Kosmala et al. (2021) [40] found a positive impact on business performance in contrast to Nenu et al. (2018) [31]. Fourthly, cash and cash equivalent to total assets have a positive impact on business performance, according to Nenu et al. (2018) [31], in contrast to García-Gómez et al. (2021) [24]. Their research shows that there is a negative relationship between return on assets and cash and cash equivalent to total assets, but this relationship is not statistically significant. Overall, while all of these studies examine the relationship between capital structure and firm performance, each study focuses on a different aspect or context, and uses different methodologies and data sources. It is important to consider the specific focus and limitations of each study when interpreting their results.

The studies listed explore the relationship between capital structure and firm performance. Results suggest that there is a significant impact of capital structure on firm performance, but the nature of the relationship may be moderated by various factors. Overall, the studies highlight the importance of considering the unique characteristics and context of each company when analyzing the relationship between capital structure and performance.

3. Methods and Materials

Sample. The total sample consists of almost 4000 small and medium-sized enterprises in the transport sector in the Central European region, such as the Czech Republic, Hungary, Poland, and Slovakia. Small and medium-sized enterprises are classified using three criteria such as sales, total assets, and employees, according to Bureau van Dijk/Moody’s Analytics [45] (see

Table 1. The classification of enterprises.

	Sales [Mil. €]	Total Assets [Mil. €]	Employees
Small company	$\left[0\right.,\left.1\right)$	$\left[0\right.,\left.2\right)$	$\left[0\right.,\left.15\right)$
Medium-sized company	$\left[1\right.,\left.10\right)$	$\left[2\right.,\left.20\right)$	$\left[15\right.,\left.150\right)$

Source: author according to Bureau van Dijk/Moody’s Analytics [45].

).

This data is drawn from Amadeus by Bureau van Dijk [45] for 2019. As can be seen,

Table 2. Sample.

		Category of Company		Total
		Small	Medium-Sized
Country	CZ	138	432	570
	HU	228	1022	1250
	PL	85	1222	1307
	SK	143	558	701
Total	V4	594	3234	3828

reveals that small businesses make up less than 16% of all businesses (594). Moreover, we find that Polish businesses form the largest group compared to others, more than one-third of all businesses. On the other hand, Czech companies make up less than 15% of all businesses (570).

Variable. 

Table 3. Capital structure variables and control variables.

Acronym	Variable	Formula	Authors	Expected Sign
TLTA	total liabilities/total assets	TL/TA	García-Gómez et al. (2021) [24], Ngatno et al. (2021) [41], Spitsin et al. (2021) [29], Nguyen and Nguyen (2020) [25], Nenu et al. (2018) [31], Vătavu (2015) [33], Vuković et al. (2022) [43], Wieczorek-Kosmala et al. (2021) [40]	−
LTLTA	long-term liabilities/total assets	LTL/TA	García-Gómez et al. (2021) [24], Tousek et. al. (2021) [42], Ngatno et al. (2021) [41], Nguyen and Nguyen (2020) [25], Nenu et al. (2018) [31], Vătavu (2015) [40], Vuković et al. (2022) [43], Wieczorek-Kosmala et al. (2021) [40]	−
CLTA	current liabilities/total assets	CL/TA	García-Gómez et al. (2021) [24], Ngatno et al. (2021) [41], Nguyen and Nguyen (2020) [25], Nenu et al. (2018) [31], Vătavu (2015) [33],	−
CACL	current assets/current liabilities	CA/CL	Tousek et. al. (2021) [42], Spitsin et al. (2021) [29], Nguyen and Nguyen (2020) [25], Nenu et al. (2018) [31], Vătavu (2015) [33], Vuković et al. (2022) [43], Wieczorek-Kosmala et al. (2021) [40], Lehenchuk et al. [44]	−
NCATA	non-current assets/total assets	NCA/TA	Spitsin et al. (2021) [29], Nguyen and Nguyen (2020) [25], Nenu et al. (2018) [31], Vătavu (2015) [33], Vuković et al. (2022) [43], Wieczorek-Kosmala et al. (2021) [40]	+
NLTA	natural logarithm of total assets	Ln of TA	García-Gómez et al. (2021) [24], Nguyen and Nguyen (2020) [25], Nenu et al. (2018) [31], Vuković et al. (2022) [43], Wieczorek-Kosmala et al. (2021) [40], Lehenchuk et al. [44]	+
CCETA	cash and cash equivalent/total assets	CCE/TA	García-Gómez et al. (2021) [24], Nenu et al. (2018) [31]	−
COM	company size	small and medium-sized company	n/a	+/−
COU	country	CZ, HU, PL, SK	n/a	+/−
LEGFO	legal form	private and public company	n/a	+/−

Note: total liabilities (TL), total assets (TA), long-term liabilities (LTL), current assets (CA), current liabilities (CL), non-current assets (NCA), cash, cash equivalent (CCE), Czech Republic (CZ), Hungary (HU), Poland (PL), and Slovakia (SK).

shows the independent indicators most commonly used to explain the relationship between capital structure and corporate performance from previous research, such as García-Gómez et al. (2021) [24], Ngatno et al. (2021) [41], Spitsin et al. (2021) [29], Tousek et al. (2021) [42], Nguyen and Nguyen (2020) [25], Nenu et al. (2018) [31] and Vătavu (2015) [33]. Finally, we extend the most frequently used indicators with categorical variables such as company size, country, and legal form. On the other hand, the most commonly used dependent variable is the return on assets, according to García-Gómez et al. (2021) [24], Ngatno et al. (2021) [41], Spitsin et al. (2021) [29], Nguyen and Nguyen (2020) [25], Nenu et al. (2018) [31] and Vătavu (2015) [33] and return on equity according to Ngatno et al. (2021) [41], Nguyen and Nguyen (2020) [25] and Vătavu (2015) [33].

These variables are summarized based on theoretical and empirical knowledge from previous studies. The capital structure is associated with several indicators such as total debt ratio, long-term debt ratio, and short-term debt ratio, as well as commonly used indicators in research on capital structure such as current ratio, non-current assets ratio, the natural logarithm of total assets or cash and cash equivalent ratio. These variables can demonstrate the effects of long-term liabilities (assets) or short-term liabilities (assets) on the capital structure.

The impact of selected financial variables on corporate performance can vary depending on the industry, region, and the specific company’s characteristics. The general overview of the expected impact of the variables is described.

• A higher TLTA indicates higher financial leverage and, thus, greater financial risk. In other words, TLTA may increase the cost of borrowing and may reduce a company’s creditworthiness. Therefore, a negative impact on corporate performance is expected.
• A higher LTLTA may indicate a higher degree of financial risk, which may decrease a company’s creditworthiness and increase the cost of borrowing. Therefore, a negative impact on corporate performance is expected.
• A higher CACL or CCETA may have a negative impact on corporate performance because free funds are not invested in development.
• A higher CLTA indicates a lower degree of liquidity and may negatively affect a company’s ability to meet its short-term obligations. Therefore, a negative impact on corporate performance is expected.
• A higher NCATA may indicate higher financial stability, which may positively impact corporate performance.
• A higher NLTA indicates a larger company size, which may positively impact corporate performance due to economies of scale.

Unfortunately, the other indicators included in our research were not found in previous studies. Among the most frequently used indicators, we also have categorical variables such as the company size according to the natural logarithm of total assets or sales; however, we classify companies according to the European Commission, further, countries such as the member states of the Visegrad Group, or the legal form broken down into trust or private companies from the transport sector.

• Company size (COM). Small and medium-sized companies may face greater challenges than larger companies due to limited resources and market power. Therefore, a negative impact on corporate performance is expected for smaller companies.
• Country (COU). A company’s location can impact its access to resources, market opportunities, and regulatory environment, among other factors. Therefore, the expected impact of a country on corporate performance will depend on the specific country’s characteristics.
• Legal form (LEGFO). Private and public companies have different reporting requirements, governance structures, and access to capital, which may impact their corporate performance. The expected impact of legal form on corporate performance will depend on the specific legal form’s characteristics.

We formulate a scientific question together with hypotheses dealing with the influence of selected factors describing the capital structure, asset structure, liquidity, and other categorical variables in the transport industry in Central Europe. What are the potentially statistically significant indicators from the capital structure together with control indicators that can be used to model business performance?

H01: 

There is a statistically significant relationship between return on assets and total liabilities/total assets.

H02: 

There is a statistically significant relationship between return on assets and long-term liabilities/total assets.

H03: 

There is a statistically significant relationship between return on assets and current assets/current liabilities.

H04: 

There is a statistically significant relationship between return on assets and current liabilities/total assets.

H05: 

There is a statistically significant relationship between return on assets and non-current assets/total assets.

H06: 

There is a statistically significant relationship between return on assets and the natural logarithm of total assets.

H07: 

There is a statistically significant relationship between return on assets and cash and cash equivalent/total assets.

H08: 

There is a statistically significant relationship between return on assets and company size.

H09: 

There is a statistically significant relationship between return on assets and country.

H10: 

There is a statistically significant relationship between return on assets and legal form.

Table 4. Descriptive statistics.

		TL/TA	LTL/TA	CA/CL	CL/TA	NCA/TA	NLTA	CCE/TA	EBIT/TA
N	Valid	3390	3407	3544	3530	3560	3828	3482	3556
	Missing	438	421	284	298	268	0	346	272
Mean		0.77	0.15	7.44	0.61	0.45	7.37	0.15	0.06
Median		0.72	0.07	1.47	0.51	0.46	7.26	0.08	0.05
Mode		0.02 a	0.00	0.12 a	0.02 a	0.00	6.64 a	−0.18 a	0.00
Std. Deviation		0.58	0.23	129.02	0.54	0.28	0.73	0.17	0.18
Skewness		5.84	4.09	45.83	5.42	−0.01	0.74	1.82	−4.84
Kurtosis		83.99	34.71	2289.44	71.32	−1.07	0.19	3.46	162.72
Range		12.20	3.82	6819.71	10.75	1.27	3.69	1.18	6.71
Minimum		−0.28	−0.17	−5.71	−0.36	−0.27	6.21	−0.18	−4.50
Maximum		11.91	3.65	6814.00	10.39	1.00	9.90	1.00	2.21
Percentiles	25	0.41	0.00	0.91	0.27	0.20	6.80	0.02	0.01
	50	0.72	0.07	1.47	0.51	0.46	7.26	0.08	0.05
	75	1.01	0.23	2.82	0.82	0.67	7.82	0.21	0.11

^a Multiple modes exist. The smallest value is shown. Note: total liabilities (TL), total assets (TA), long-term liabilities (LTL), current assets (CA), current liabilities (CL), non-current assets (NCA), cash, and cash equivalent (CCE).

shows that transport companies have an average debt ratio of 77% in V4; the median is lower. Moreover, descriptive statistics demonstrate that 75% of businesses have at least a 41% debt ratio. In other words, these results indicate that companies use foreign capital relatively more often than their resources. We find that the 77% debt ratio consists of 15% long-term debt, and the rest is short-term debt because the average long-term debt ratio is 15% and the average short-term debt ratio is more than 60%. However, the median long-term debt ratio is more than two times lower, as the median long-term debt ratio is 7%. On the other hand, the difference between the average and the median long-term debt ratio is approximately 10%. The results also show that 25% have a short-term debt ratio of at least 82%. The transport sector has a significantly different mean and median current ratio, as the average business has more than seven times more current assets than current liabilities, but the median is less than 1.50. In addition, we find that companies from the transport sector have, on average, 15% of their assets in cash or cash equivalents; the median is almost two times lower than the average. Another indicator of the structure of assets demonstrates that non-current assets, on average, make up 45% of all assets; the median is about 1% higher. The company size was calculated as the natural logarithm of the total assets, indicating that the indicator ranges from 6.21 to 9.90, and the average size of the company is more or less similar to the median. Finally, descriptive statistics indicate that the company’s performance on average is 6% in the transport sector, but this indicator varies from (−) 450% to 221%; even 25% of companies have a return on assets of less than 1%.

Table 5. Descriptive statistics (without outliers).

		TL/TA	LTL/TA	CA/CL	CL/TA	NCA/TA	NLTA	CCE/TA	EBIT/TA
N	Valid	3319	3257	3130	3424	3560	3784	3251	3235
	Total	509	571	698	404	268	44	577	593
	Missing	438	421	284	298	268	0	346	272
	Outliers	71	150	414	106	0	44	231	321
Mean		0.72	0.12	1.66	0.56	0.45	7.34	0.11	0.05
Median		0.70	0.06	1.33	0.50	0.46	7.25	0.07	0.04
Mode		0.02 a	0.00	0.12 a	0.02 a	0.00	6.64 a	−0.18 a	0.00
Std. Deviation		0.41	0.14	1.18	0.37	0.28	0.69	0.12	0.07
Skewness		0.37	1.15	1.26	0.67	−0.01	0.59	1.18	0.41
Kurtosis		−0.35	0.38	1.25	−0.12	−1.07	−0.26	0.60	0.26
Range		2.17	0.74	5.66	2.00	1.27	3.15	0.66	0.40
Minimum		−0.28	−0.17	0.00	−0.36	−0.27	6.21	−0.18	−0.14
Maximum		1.89	0.57	5.66	1.64	1.00	9.36	0.47	0.25
Percentiles	25	0.41	0.00	0.84	0.26	0.20	6.79	0.02	0.01
	50	0.70	0.06	1.33	0.50	0.46	7.25	0.07	0.04
	75	0.99	0.21	2.18	0.79	0.67	7.80	0.17	0.09

^a Multiple modes exist. The smallest value is shown. Note: total liabilities (TL), total assets (TA), long-term liabilities (LTL), current assets (CA), current liabilities (CL), non-current assets (NCA), cash, and cash equivalent (CCE).

reveals descriptive statistics without outliers of selected financial indicators for the final sample of all transport enterprises in V4. This descriptive statistic provides relevant results unlike the previous one; for example, the mean and median current ratio achieves a more or less similar result. Twenty-five percent of enterprises have over two times more current assets than short-term liabilities. The capital structure consists of 72% liabilities on average, and the median is 2% smaller. The results show that companies prefer short-term liabilities to long-term liabilities based on the average and median long-term debt and short-term debt ratio. Furthermore, the company has on average, cash and other equivalents of around 11%; 25% of transport companies have at least 17% of all assets in cash and other equivalents, and the current ratio ranges from (−) 0.18 to 0.47. In addition, we find that the company size ranges from 6.21 to 9.36. Finally, corporate performance averages around 5%, and the median is about 1% lower. This indicator ranges from (−) 14% to 25%.

Methods. We use multiple linear regression to identify the relationship between the debt ratio and independent variables. This method is a supervised machine learning algorithm containing a broader data spectrum. The purpose of multiple regression is to predict dependent variables based on potential input variables. The main output is a formula explaining the factor impact on the dependent variable. Multiple linear regression requires assumptions such as homogeneity of variance (homoscedasticity), no multicollinearity, independence of observations, normality, and linear relationship between dependent and independent variables. The model selects statistically significant variables using backward elimination, forward selection, and stepwise selection. We apply stepwise selection. This method combines forward and backward processes. The stepwise method is a procedure that attempts to find the best multiple linear regression model, including statistically significant variables from a more extensive set of potential variable datasets.

$$\mathrm{y}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{X}}_1+\cdots +{\mathsf{\beta }}_{\mathrm{n}}{\mathrm{X}}_{\mathrm{n}}+\mathsf{\epsilon }$$

(1)

where

$\mathrm{y}$→dependent variable,

${\mathsf{\beta }}_0$→intercept,

${\mathsf{\beta }}_1$→the regression coefficient of the first independent variable,

${\mathrm{X}}_1$→the first independent variable,

${\mathsf{\beta }}_{\mathrm{n}}$→the regression coefficient of the last independent variable,

${\mathrm{X}}_{\mathrm{n}}$→last independent variable,

$\mathrm{n}$→the number of predictor variables,

$\mathsf{\epsilon }$→model error.

Outliers. Mahalanobis distance is a measure of the distance between two points in a multivariate space. It considers the correlations between variables and measures the distance in terms of standard deviations. It is used in various statistical applications such as pattern recognition, cluster analysis, and outlier detection. The Mahalanobis distance is the square root of the chi-squared statistic and is a more robust metric compared to the Euclidean distance.

$${\mathrm{D}}^2={\left(\mathrm{x}-\mathrm{m}\right)}^{\mathrm{T}}{\mathrm{C}}^{-1}\left(\mathrm{x}-\mathrm{m}\right)$$

(2)

where

${\mathrm{D}}^2$→Mahalanobis distance,

$\mathrm{x}$→vector or data,

$\mathrm{m}$→vector of the mean of independent variables,

${\mathrm{C}}^{-1}$→the inverse covariance matrix of independent variables,

$\mathrm{T}$→indicates that the vector should be transposed.

Multicollinearity is a statistical concept that refers to a high degree of correlation among predictor variables in a regression analysis. When two or more predictors are highly correlated, they provide redundant information and can lead to unstable and misleading results. This is because the predictors are not independent, and small changes in the data can result in large changes in the estimated coefficients. Multicollinearity can affect the validity of hypothesis tests, confidence intervals, and prediction accuracy. It can be diagnosed by examining the variance inflation factor (VIF) or by calculating the correlation matrix among the predictors. To address multicollinearity, one can remove one of the highly correlated predictors or combine them into a single composite predictor.

The VIF is a measure of how much the variance of a regression coefficient is increased due to multicollinearity. A VIF score of 1 means there is no multicollinearity, while a score greater than 1 indicates that the predictor is highly correlated with other predictors in the model. A VIF greater than 5 or 10 is generally considered high.

$$\mathrm{V}\mathrm{I}\mathrm{F}=\frac{1}{1-{\mathrm{R}}_{\mathrm{i}}^2}=\frac{1}{\mathrm{t}\mathrm{o}\mathrm{l}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}}$$

(3)

where

$\mathrm{V}\mathrm{I}\mathrm{F}$→variation inflation factor,

${\mathrm{R}}_{\mathrm{i}}^2$→unadjusted coefficient of determination for regressing the ith independent.

The Durbin-Watson statistic is a statistical test used to detect the autocorrelation in the residuals of a regression model. Autocorrelation occurs when the residuals of a regression model are not independently distributed but instead are correlated with each other. This can cause the regression coefficients to be inefficient and may lead to incorrect hypothesis tests and confidence intervals.

The Durbin-Watson statistic is a number between 0 and 4, where values close to 2 indicate that there is no autocorrelation, while values close to 0 or 4 indicate positive or negative autocorrelation, respectively. If the Durbin-Watson statistic is close to 2, it suggests that the residuals are independently distributed and the regression model is well-specified. If the Durbin-Watson statistic is not close to 2, it suggests that there may be autocorrelation in the residuals, and the model may need to be modified to address this issue.

$$\mathrm{D}\mathrm{W}=\frac{\sum _{\mathrm{t}=2}^{\mathrm{T}}{\left({\mathrm{e}}_{\mathrm{t}}-{\mathrm{e}}_{\mathrm{t}-1}\right)}^2}{\sum _{\mathrm{t}=1}^{\mathrm{t}}{\mathrm{e}}_{\mathrm{t}}^2}$$

(4)

where

${\mathrm{e}}_{\mathrm{t}}$→residual figure,

$\mathrm{T}$→number of observations of the experiment.

The coefficient of determination, often denoted as R², is a statistic that measures the proportion of variation in the dependent variable that is predictable from the independent variable(s) in a regression model. It ranges from 0 to 1, where 0 indicates that the model does not explain any of the variations in the dependent variable, and 1 indicates that the model explains all of the variations. A high R² value indicates that the model fits the data well and can be used to make reliable predictions.

The adjusted R-squared is a modification of the R-squared that adjusts the statistic based on the number of predictors in the model. It adjusts the R-squared value by penalizing the inclusion of additional predictors that do not improve the model fit. The adjusted R-squared ranges from 0 to 1, just like the R-squared, but it gives a more accurate estimate of the model’s ability to explain the variation in the dependent variable, especially when comparing models with different numbers of predictors. A high adjusted R-squared value indicates that the model fits the data well and includes only the important predictors.

$${\mathrm{R}}^2=1-\frac{\mathrm{S}\mathrm{S}\mathrm{R}}{\mathrm{S}\mathrm{S}\mathrm{T}}$$

(5)

$$\mathrm{a}\mathrm{d}\mathrm{j}.{\mathrm{R}}^2=1-\left(\frac{\mathrm{n}-1}{\mathrm{n}-\left(\mathrm{p}+1\right)}\right)\left(1-{\mathrm{R}}^2\right)$$

(6)

where

SSM→sum squared regression,

SST→total sum of squares,

${\mathrm{R}}^2$→sample R-squared,

$\mathrm{n}$→the sample size,

$\mathrm{p}$→number of the independent variable.

Figure 1 shows that all prerequisites for applying linear regression analysis are met. This method requires normal distribution and homoscedasticity.

4. Result

Table 6. Model summary.

Step	R	R Square	Adjusted R Square	Std. Error of the Estimate	Change Statistics		df1	df2	Sig. F Change	Durbin-Watson
					R Square Change	F Change
1	0.257 a	0.066	0.066	0.069	0.066	153.409	1	2173	0.000
2	0.291 b	0.085	0.084	0.068	0.019	44.675	1	2172	0.000
3	0.326 c	0.106	0.105	0.067	0.021	51.498	1	2171	0.000
4	0.351 d	0.123	0.122	0.066	0.017	42.975	1	2170	0.000
5	0.356 e	0.127	0.125	0.066	0.003	8.178	1	2169	0.004
6	0.360 f	0.129	0.127	0.066	0.003	6.547	1	2168	0.011	1.902

^a Predictors: (Constant), CCETA. ^b Predictors: (Constant), CCETA, HU. ^c Predictors: (Constant), CCETA, HU, Medium. ^d Predictors: (Constant), CCETA, HU, Medium, CACL. ^e Predictors: (Constant), CCETA, HU, Medium, CACL, TLTA. ^f Predictors: (Constant), CCETA, HU, Medium, CACL, TLTA, NCATA. Dependent Variable: EBITTA.

shows that the model explains almost 13% of the variability of the output variable, namely return on assets calculated as the proportion of EBIT to total assets, using a statistically significant model consisting of six significant independent variables. These variables are selected using the stepwise method. As can be seen, the process consists of six steps; each step includes a different statistically significant variable, such as CCETA, HU, Medium, CACL, TLTA, and NCATA. The table shows a universal model for all transport companies. This model was created in six steps, each step identifies a statistically significant variable, and we use the stepwise method.

We find that the model consists of two categorical variables (see

Table 7. Multivariate linear regression model ^a.

	Unstandardized Coefficients		t	Sig.	95.0% C. I. for B		Collinearity Statistics
	B	Std. Error			Lower Bound	Upper Bound	Tolerance	VIF
(Constant)	0.025	0.010	2.487	0.013	0.005	0.045
CCE/TA	0.113	0.016	7.320	0.000	0.083	0.144	0.731	1.368
HU	0.023	0.003	7.528	0.000	0.017	0.029	0.915	1.093
Medium	0.030	0.004	7.154	0.000	0.022	0.039	0.969	1.032
CA/CL	0.004	0.002	2.043	0.041	0.000	0.008	0.397	2.520
TL/TA	−0.021	0.006	−3.672	0.000	−0.033	−0.010	0.486	2.060
NCA/TA	−0.017	0.007	−2.559	0.011	−0.030	−0.004	0.644	1.552

^a Dependent Variable: EBIT/TA.

), such as the company size divided into small and medium-sized enterprises according to Bureau van Dijk [45] and the country divided into the Czech Republic, Hungary, Poland, and Slovakia. As can be seen, the company size shows that a medium-sized company in the transport industry has a higher ROA than a small company, assuming other variables are unchanged. Moreover, the model demonstrates that Hungarian transport companies are more profitable than companies from other countries of the Visegrad Group. This indicator has a similar coefficient to the coefficient for a medium-sized enterprise; the result reveals that Hungarian medium-sized enterprises are more profitable than others, assuming that other variables are unchanged. Finally, we find that the legal form has no statistically significant impact on corporate profitability; the legal form categorized the companies into private and public companies.

Four of the six significant indicators are numerically describing capital structure, asset structure, or liquidity. VIF emphasizes that there is no multicollinearity between all indicators; all VIFs are less than five. These indicators were selected based on theoretical and empirical knowledge from previous research. The model demonstrates that two variables such as TL/TA and NCA/TA, have a negative impact on the corporate performance of transport companies. If the debt ratio (TLTA) rises, profitability will fall. Likewise, if the non-current assets to total assets (NCATA) rise, profitability will again fall. In other words, transport companies should not prefer foreign resources to their resources; this capital should be allocated to current assets. Moreover, we find that liquidity (CACL) has a positive impact on corporate performance, similar to the cash and other cash equivalents to total assets (CCETA). Other variables such as LTL/TA, CL/TA, or NLTA have no statistically significant impact on business performance.

5. Discussion

Table 8. Summary.

Authors	Data	Country	Sample	Sector	Dependent Variable	Coefficient	Sig.	Independent Variable	Method	R-Square	adj. R-Square
Vuković et al. (2022) [43]	2013–2019	EU	460	agriculture	ROA	3.249		constant	MLR	0.305	n/a
						−11.908	***	STLR
						−21.590	***	LTLR
						−2.281	***	TL/TA
						0.000		TL/TE
						1.539	***	log of TA
						−6.781	***	net FA/TA
						0.086	***	CA/CL
Lehenchuk et al. (2022) [44]	2020	SK	88	advertising	ROA	−1.074	***	constant	OLS	0.953	0.951
						0.003		CA/CL
						−0.004	*	TL/TE
						0.958	***	TS/TA
						0.074	**	ln of TA
Wieczorek-Kosmala et al. (2021) [40]	2015–2019	HU, PL, SK, CZ	1977	energy	ROA	−1.506	***	constant	WLS	0.449	0.447
						−0.321	***	TL/TA
						−0.102	***	age
						0.070		ln of TA
						−0.166	***	sector
						0.000		FA/TA
						−0.138	***	CA/CL
						−0.092	***	financial slack
						0.124	***	TS/TA
						0.561	***	EBIT/TS
					ROA	−1.951	***	constant	WLS	0.753	0.752
						−0.319	***	LTL/TA
						0.209	***	age
						0.069	***	ln of TA
						0.178	***	sector
						−0.086	***	FA/TA
						0.248	***	CA/CL
						−0.227	***	financial slack
						0.012		TS/TA
						0.547	***	EBIT/TS
					ROA	−2.734	***	constant	WLS	0.662	0.661
						0.092	**	STL/TA
						−0.187	***	age
						0.369	***	ln of TA
						0.042	**	sector
						−0.354	***	FA/TA
						−0.114	**	CA/CL
						0.177	***	financial slack
						0.028		TS/TA
						0.653	***	EBIT/TS
our research	2019	CZ HU PL SK	3828	transport	ROA	0.025	**	constant	MLR	0.129	0.127
						−0.021	***	TL/TA
								LTL/TA
						0.004	**	CA/CL
								CL/TA
						−0.017	**	FA/TA
								NLTA
						0.113	***	CCE/TA
						0.023	***	HU
						0.030	***	medium-sized company
								LEGFO

Note: short-term liability ratio (STLR), long-term liability ratio (LTLR), total liabilities (TL), total assets (TA), total equity (TE), fixed assets (FA), current assets (CA), current liabilities (CL), total sales (TS), long-term liabilities (LTL), short-term liabilities (STL), earnings before interest and tax (EBIT), multivariate linear regression (MLR), weight least square (WLS), linear regression (LR). ROA is a dependent variable calculated by EAT/TA in all previous research, but our research uses EBIT/TA. p-value < 0.01 (***), p-value < 0.05 (**), p-value < 0.10 (*).

compares our outputs with the results of previous research on the influence of selected indicators describing primarily the capital structure on the corporate performance of European companies, especially companies from the Central European area. This research is suitable for comparison with our outputs because the companies are from similar regions. The table includes research for various industries, such as agriculture, advertising, and the energy industry. We summarize several data about this research, such as period, sample size, country, and industry, but especially the coefficients forming the performance estimation model together with the method used. Finally, we confront our models with other models using an adjusted R-square. The results show that firm size calculated as the natural logarithm of total assets is a statistically significant variable in all models except one by Wieczorek-Kosmala et al. (2021) [40]. Their models differ because the first model contains a debt ratio, but the other models contain a long-term debt ratio or a short-term debt ratio. These models demonstrate that asset size has a positive impact on increasing corporate performance. On the other hand, a higher proportion of non-current assets to total assets reduces business performance. In other words, these models estimate that current assets are an important pillar in managing business performance. Other indicators such as CA/CL, TL/TE, or TS/TA differ in their impact on business performance because the current ratio (CA/CL) is a statistically significant variable according to Vuković et al. (2022) [43], Wieczorek-Kosmala et al. (2021) [40], unlike Lehenchuk et al. (2022) [44]. In addition, the current ratio provides different findings because the impact is not clear-cut since the coefficient is sometimes negative and sometimes positive. Furthermore, TL/TE has a negative statistically significant impact on business performance, according to Lehenchuk et al. (2022) [44]; however, the regression coefficient is very low. On the other hand, Vuković et al. (2022) [43] did not find a relationship between return on assets and TL/TE; however, TS/TA has a significant positive impact on the possible increase in business performance according to Lehenchuk et al. (2022) [44], but also one of the three models from Wieczorek-Kosmala et al. (2021) [40]. Other factors are compared, and other researchers have used them. Moreover, these models are not the only ones in estimating corporate performance; Norvaisiene (2012) [46] models performance for companies from Latvia, Estonia, and Lithuania outside Europe, such as Saudi Arabia Ghardallou (2022) [47] or Mongolia Bayaraa, (2017) [48].

Our results are compared with previous results. We find that our model, like other models by Vuković et al. (2022) [43] and Wieczorek-Kosmala et al. (2021) [40], emphasizes the negative impact of debt ratio on corporate performance. In our model, this coefficient does not have a significant impact, unlike other models. On the other hand, the current ratio has a positive impact on the profitability of assets, unlike Wieczorek-Kosmala et al. (2021) [40] except for one of the four models and Vuković et al. (2022) [43]. Other indicators, such as company size calculated as the natural logarithm of total assets and long-term debt ratio, do not have a statistically significant impact on company performance, unlike Vuković et al. (2022) [43], Lehenchuk et al. (2022) [44] or Wieczorek-Kosmala et al. (2021) [40]. Tangibility does not differ from other models; this coefficient reduces business performance similar to other sectors, such as agriculture and energy, according to Vuković et al. (2022) [43] and Wieczorek-Kosmala et al. (2021) [40]. Finally, indicators such as CLTA, CCE/TA, and categorical variables such as country, company size according to Dijk’s classification, and legal form are not compared because these indicators are not part of the output models.

Finally, we summarise that TL/TA and NCA/TA have a negative impact on the profitability of transport companies. This means that transport companies should avoid relying too much on debt and non-current assets and should instead focus on allocating their capital to current assets. On the other hand, liquidity indicators such as CACL and CCETA have a positive impact on corporate performance. This indicates that transport companies should maintain a sufficient level of liquidity to ensure their financial stability and ability to meet their obligations. Lastly, the analysis suggests that other variables, such as LTL/TA, CL/TA, or NLTA, do not have a statistically significant impact on business performance. This means that transport companies can focus on other areas to improve their performance, such as increasing revenue or reducing costs. Overall, this analysis provides valuable insights for transport companies in managing their financial resources and optimizing their performance.

6. Conclusions

We aimed to point out the impact of the capital structure, expressed mainly using selected indicators, on business performance in the transport sector in the Visegrad Group using multiple regression analysis. We found that six out of ten indicators have a statistically significant impact on the business performance of transport companies from Central Europe. The results show that the debt ratio and the non-current assets ratio have a negative impact on profitability, as both of these variables show a negative coefficient. In other words, companies with a high debt ratio and a high non-current assets ratio reduce corporate performance. On the other hand, the current ratio, cash, and its equivalent to total assets have an unusually positive impact on profitability. In general, high liquidity, or the assets with the highest level of liquidity, do not normally have a positive impact on performance because these funds are not invested in business development. Finally, we found that the company should consider the location of the company headquarters because Hungarian transport companies have more potential to achieve performance in contrast to companies in other countries. Moreover, if investors are considering a potential investment in a small or medium-sized business, investing in a medium-sized business is better than a small business. Hungarian medium-sized enterprises have the best prerequisite for achieving profitability from the point of view of these two categorical variables. The resulting model explains almost 13% of the variability of the return on assets using six statistically significant variables. These indicators were selected using the stepwise method. Comparing our model with other models from different industries, we find that the indicators most often used from a wide range of theoretical and empirical knowledge describing the impact of capital structure on business performance do not play a significant role in the transport industry, unlike agriculture, energy or the marketing industry based on adj. R-square. These results benefit businesses, investors, and other stakeholders in modeling business performance for a specifically different industry in Central Europe as opposed to others. Managers of transport companies should prefer equity over external liabilities; in addition, high liquidity helps to increase business performance.