Assessment of bankruptcy risks in Czech companies using regression analysis

Bankruptcy is an important topic in academic research and practice. It is a burning issue worldwide in the current COVID-19 situation. The aim of this study is to examine the financial risks of Czech companies. By employing the stepwise regression technique to estimate the financial risks, the p-values of all selected 15 financial ratios (explanatory variables) were calculated. If the p-value of the variable is more than the level of significance, the particular variable is removed from the model and another regression model is calculated. The findings from the stepwise regression revealed that return on capital, current ratio, net working capital turnover rate, and current assets turnover rate have a positive influence on company’s financial health. On the contrary, return on capital employed, asset turnover rate, inventory turnover rate, fixed assets turnover rate, and debt to equity ratio negatively impact the company’s financial health. The findings of this study will be fruitful for managers, policymakers, and investors of the companies to estimate and assess financial risks.
AcknowledgmentsThis study is supported by the Internal Grant Agency (IGA) in Tomas Bata University in Zlin, the Czech Republic, under the projects No IGA/FAME/2021/008 and IGA/FAME/2021/014.


INTRODUCTION
The consequences of the COVID-19 pandemic could be seen and felt around the world. European countries, mainly Central European countries, are also being intensely affected by the pandemic. The pandemic is severely affecting different sectors of the economies, labor markets, and companies, becoming a cause of bankruptcy for several famous brands in many industries (Donthu & Gustafsson, 2020). However, some companies and businesses are struggling, and some are booming in this situation. According to the Cribis database statistics, around 16,111 companies were shut down in the Czech Republic in 2020, which was the highest number in Czech history in a year. About 1,336 companies in January 2021, 1,093 companies in February 2021, and 1,260 companies in March 2021 were shut down in the country. Therefore, bankruptcy has become a burning issue in the current situation.
Assessment of bankruptcy risks are important not only for investors that consider making investment decisions but also for managers and policymakers to make decisions on how to improve the financial performance of the companies. Kristóf and Virág (2020) argued that bankruptcy prediction has gone an important consideration in the last many decades, but the prediction in ex-socialist countries got substantial attention more than 20 years later compared to western countries. Therefore, Kristóf  The uniqueness of the current study is that it uses the financial ratios that can be used in the prediction model for Czech enterprises. This study is highly relevant as it focuses on the assessment of bankruptcy risks of Czech companies using the financial ratios of multiple sectors simultaneously to a comprehensive overview. The main aim of this study is to examine the financial risks of Czech companies. Therefore, liquidity, activity, profitability, and indebtedness ratios were comprised as these ratios affect the financial health of the companies.
There are different methods and models used to estimate business failure over time such as Altman (1968), Springate (1978), Ohlson (1980), and Zmijewski (1984) models, etc. Bărbuță-Mișu and Madaleno (2020) claimed that these models use different statistical techniques and independent variables, and offer valuable information about the financial risks and financial performance of the companies. However, Valaskova et al. (2018a) and Valaskova et al. (2018b) mentioned and applied the decisive criteria at the Slovak companies to examine the financial risk of the Slovak companies. Slovakia and the Czech Republic are countries with similar history, political, economic, social, and legislative environments; therefore, it can be assumed that the same criteria can be used in Czech conditions. Therefore, the decisive criteria are employed to examine the bankruptcy risks in Czech companies. The criteria are that if a company's net income is negative, equity to debts ratio is less than 0.08, and current ratio is less than 1, then the company is in default or unhealthy. The outcomes were obtained of the selected explanatory variables by using the stepwise regression analysis.

LITERATURE REVIEW
Any risk is, in fact, the likelihood that adverse conditions will occur or that an adverse event will occur. Risks associated with financial activities are called financial risks. This term refers to the probability that the actual return will be less than predicted. Knowing the financial risk is very important for investors and stakeholders, especially when considering future decisions. Depending on the level of risk, investors decide how much of their free funds to invest in a given commodity. Risk-averse investors prefer the lowest possible risk, even at the cost of lower profits. The problem of predicting the financial risks of a company has become a crucial area of research. Therefore, scholars from all over the world are engaged in developing forecasting models to predict financial risks.

DATA AND METHODOLOGY
The secondary data of the Czech companies was obtained from the CRIBIS database, where the data is classified according to CZ-NACE. CRIBIS is a part of the global CRIF group, which was founded in 1988. Cepel  In the first step, 15 financial ratios (independent variables) of the Czech companies in the selected sample were computed. The complete detail about the selected variables (ratios) and their measurement is given in Table 2. Based on the computed financial ratios, the companies were classified into two groups by the decisive criteria: healthy or non-default companies; and unhealthy or default companies. If a company has positive net income, the ratio (X6) is greater than 1, and the equity to debts ratio is more than 0.08, then the company is healthy or non-default (marked by the value 1). If these conditions are not met, the company is unhealthy or default (marked by the value 0). According to the calculation of the decisive criteria, Czech companies are classified into bankrupt and non-bankrupt firms. The sample consists of 7,779 Czech companies from 11 sectors, where around 13.1% of them experienced some financial risks or are unhealthy. Table 1 provides information about the Czech companies by sector-wise data of total and unhealthy companies. Table 1 was prepared after em-ploying the decisive criteria. Table 1 revealed that sector 5 has only 1.43% of total companies by comparing to other sectors, and the default percentage of sector 5 is 0.98%, which is the lowest percentage among default companies. On the other hand, sector 11 has 29.58% of companies in the current study, and the percentage of default companies is the highest than other sectors. However, the proportion of default companies is higher than the total number of companies in sectors 1, 4, and 8. Conversely, the proportion of default companies is lower than the total number of companies in sectors 3, 7, and 11. According to Table 2, there are two types of variables in the regression analysis: independent variables (15 financial ratios) and dependent variables (financial health of the company). All the data of the financial ratios were treated in linear regression analysis. As the independent variable has 15 financial ratios, multiple regression has been employed to find out which financial ratio has an impact on the financial health of the company. The financial ratios such as profitability ratios, turnover ratios, activity ratios, liquidity ratios, and financial structure were calculated for the selected Czech companies. Equation 1 is presented the linear regression model. The main objective of the regression analysis is to find the existing independence and to examine the relationship, which changes one variable dependence on others. As the number of explanatory variables is more than one, multiple regres-

RESULTS
The descriptive statistics of all the selected variables is presented in Table 3. All estimations were performed using STATA 16.0 software.
According to Table 3, the mean and standard deviation of the selected variables are slightly different from each other. The values of mean, standard deviation, minimum, and maximum of X 8 are different from all other variables. The reason is that the numerator (earnings before interest and taxes) of X 8 is much bigger than its denominator (interest expense). In Table 3, the values of kurtosis and skewness are computed to check the normality in the data. For the standard normal distribution, it is recommended that the value of kurtosis should be zero, which is unlikely for real-world data. For the normal distribution, Simon et al. (2017) suggested that the skewness and kurtosis should be within the range ± 3, and ± 10, respectively. According to Table 3, most of the values of skewness and kurtosis are showing that the data used in the study is normally distributed. To check the multicollinearity in the data, the variance inflation factor (VIF) coefficient criteria was employed. VIF coefficient is the indicator that diagnoses the existence of multicollinearity between the independent variables (Sharma et al., 2020). Multicollinearity is not a serious problem if the value of the VIF coefficient of each explanatory variable is less than 10 (Nachane, 2006). Table 4 is describing the values of the VIF coefficient and 1/ VIF. All the values of VIF coefficients of the independent variables are less than 10, therefore, there is no multicollinearity issue in the data. The data can be processed for further analysis.
In the multiple regression analysis, the p-values of all 15 financial ratios (variables) were calculated to compare with the level of significance. If the p-value of a financial ratio is more than 0.05 (level of significance), then that particular financial ratio was removed from the model and another regression was calculated. Therefore, the stepwise regression was used to recognize the least significant financial ratio in the model to realize which factors are significant enough to manage financial risks and to forecast the default of the Czech companies. Table 5 shows the outcomes of the first regression model for all 15 financial ratios. From Table 5, it is clear that the p-value of X 6 is the highest of other values. Therefore, X 6 is removed from the model and the regression is run again. Table 6 shows the results of the regression, as well as the p-value of all the financial ratios. According to Table 6, it is clear that the p-value of X 1 (return on assets) is the highest of other values. Therefore, X 1 is removed from the model and the regression is run again. Table 7 shows the results of the regression. The p-value of all the independent variables is less than the value of significance (0.05). After employing the stepwise regression, the results are presented in Table 7. All the selected variables (financial ratios) are statistically insignificant as the p-values of all the independent variables are less than 0.05.

DISCUSSION
From Table 7, it can be observed that the following ratios have significant impacts on the financial health of Czech companies. The significant financial ratios are: return on capital employed (X 3 ), return on capital (X 4 ), current ratio (X 5 ), interest coverage ratio (X 8 ), inventory turnover rate (X 10 ), asset turnover rate (X 11 ), net working capital turnover rate (X 12 ), fixed assets turnover rate (X 13 ), current assets turnover rate (X 14 ), and debt to equity ratio (X 15 ). Hence, the model can be presented like this in equation 2.
The findings from the stepwise regression in equation 2 revealed that return on capital employed (X 3 ), inventory turnover rate (X 10 ), asset turnover rate (X 11 ), fixed assets turnover rate (X 13 ), and debt to equity ratio (X 15 ) have a negative impact on the dependent variable. The Czech companies should decrease the values of these ratios to increase the financial health of their companies, holding all other variables constant. Increasing one unit in X 3 , X 10 , X 11 , and X 15 will raise the risk of the financial health of the Czech companies by 0.024, 0.001, 0.051, and 0.040 units, respectively. The magni-tude of the coefficient of X 10 (inventory turnover rate) is very small, conversely, the magnitude of X 11 (asset turnover rate) is big among financial ratios that have a negative impact on the financial health of the Czech companies. On the other hand, return on capital (X 4 ), current ratio (X 5 ), interest coverage ratio (X 8 ), net working capital turnover rate (X 12 ), and current assets turnover rate (X 14 ) have a positive influence on the financial health of the companies. Therefore, the Czech companies should increase the values of return on capital employed, current ratio, net working capital turnover rate, and current assets turnover rate. In this way, the Czech companies will increase their financial health, holding all other variables constant. Increasing one unit in X 4 , X 5 , X 8 , X 12 , and X 14 will increase the financial health of the Czech companies by 0.295, 0.057, 0.000, 0.001, and 0.022 units, respectively. The magnitude of the coefficient of X 4 (return on capital employed) is big, conversely, the magnitude of X 8 (interest coverage ratio) is very small among financial ratios that have a positive impact on the financial health of the Czech companies. R 2 is known as the coefficient of determination that indicates how close the data are to the fitted regression line. The value of R 2 is 68.3 % and adjusted R 2 is 68.2%. The value of Prob > F shows the significance of the model as the value is lower than the significance level (0.05). Hence, the formed regression model of the bankruptcy prediction of the Czech companies is statistically significant.

CONCLUSION
The assessment of bankruptcy risks is essential for investors that consider making investment decisions on bonds, equity, and creditors. The assessment of the bankruptcy risks is also significant for managers and policymakers as they make decisions to improve the financial performance of the companies. The current study provides a comprehensive review regarding the empirical research of bankruptcy prediction of Czech companies. The main aim of this study is to investigate the financial risks of Czech companies.
To explore the aim, the secondary data of the Czech companies was obtained from the CRIBIS database for the period of 2019. The final sample contained 7,779 Czech companies from eleven sectors. In the multiple regression analysis, p-values of the selected financial ratios were calculated to compare with the level of significance (0.05). If the p-value of a financial ratio is more than 0.05, then the financial ratio is removed from the model, and another regression model is estimated by using the stepwise regression technique. Finally, the statistically significant ratios that affect the future financial development of the company were estimated. The outcomes through multiple regression analysis exposed that return on capital employed, inventory turnover rate, asset turnover rate, fixed assets turnover rate, and debt to eq-