“Impact of intellectual capital on profitability: Evidence from software development companies in the Slovak Republic”

Intellectual capital is the total value of all entity’s intangible resources (organization- al, human, and customer). Effective management of intellectual capital in high-tech industries needs determination of its role in ensuring profitability and clarifying the direction of managerial and investment policy in intangible resources. The aim of this study is to investigate the impact of intellectual capital on the profitability of Slovak software development companies. Panel data regression analysis was used as the main research method to analyze the data of 16 Slovak software development companies for 2015–2019. The study designed and analyzed four panel data regression models with different dependent variables (Return on Assets, Net Profit Margin, Gross Profit Margin, Earnings Before Interest and Taxes Margin) and similar independent variables (Capitalized Development Costs, Software, Acquired Intangible Fixed Assets, Personnel Costs, Social Security Costs, Social Costs, and Total Costs of Economic Activity). The analysis of these models was carried out based on the fixed effects meth- od. It was found that intellectual capital reflected in the financial statements of software development companies does not meet the information needs of stakeholders and does not have a significant direct impact on profitability. Only Acquired Intangible Fixed Assets had a direct positive impact on the profitability of software development companies in all four analyzed models, and some independent variables had a negative impact. It is proposed to expand the structure of financial reporting items that charac- terize the intellectual capital and improve the method of recognizing costs of various types as intangibles. role


INTRODUCTION
The formation of the knowledge economy in a world is characterized by changing the priority of resources used by enterprises to ensure profitability. This has led to special attention to intellectual capital management as a significant factor in value creation. The role of intellectual capital in ensuring profitability is obvious for innovative enterprises, which are the main driving force of economic development. Many countries around the world support the development of innovative sectors of the economy, in particular, software development companies.
The shift in management's focus from tangible to intangible capital when considering the 'value creation' processes within firms (Abeysekera, 2008) also raised a number of important questions among scientists regarding the need to improve the theoretical and methodological foundations of intangible assets management. The current financial statements of innovative enterprises do not fully meet the needs of management in the operation of intellectual capital. From the standpoint of representatives of the theory of post-industrial society and the knowledge economy (Adams & Oleksak, 2010;Haskel & Westlake, 2018;Kabir, 2019;Ullberg et al., 2021), primarily the innovative enterprises must have the strongest link between their intellectual capital or intangibles and the level of their profitability. Therefore, the imperfection of the current financial reporting primarily also has negative consequences for the management of software development companies and their capital suppliers due to the inability to provide relevant information about intellectual capital.
This study tests whether the profitability of innovative enterprises not significantly depends on the available intellectual capital reflected in financial statements. Therefore, it can serve as empirical evidence for the improvement of managerial and investment policy of software development companies in intangible resources. In addition, it highlights the necessity of improving current information provision of intellectual capital management, taking into account the sectoral characteristics of enterprises and the specifics of the use of intangibles.

LITERATURE REVIEW
The growth of profitability is the defining goal of any enterprise, so the analysis of factors influencing profitability is the research focus of many scholars. Analyzed papers in this area in the context of this study can be divided into two main groups. The first group presents studies that conduct a general analysis of the impact of various factors of intangible nature (intangible assets, intellectual capital, intellectual resources, investment in intangible assets, etc.) on profitability. The second group includes impact analysis of intellectual capital directly on the profitability of high-tech companies (IT companies, software developers).

Impact of intangible factors on the profitability management
The influence of intangible factors on profitability was considered both directly and within the frame of their impact on financial performance because profitability is one of its components. This allowed obtaining a number of important results that explain the role of various types of intangible factors in ensuring the profitability of companies in different sectors of the economy.
From a theoretical point of view, intellectual capital should positively affect the profitability of enterprises and should play a decisive role among all other factors. For instance, Sedláček (2010) investigated the role of intangible assets that have not been included in the balance sheet of the Slovak and Czech companies' market value. It was found that such assets represent the main item of the estimated market value of companies, ensuring their profitability. Yuan and Rizki (2020) continued the development of the idea of taking into account the unrepresented part of intellectual capital in analyzing its impact on profitability. They pointed out that intangible assets as unexplained value have a significant and positive effect on the financial performance of manufacturing companies listed on the Indonesia Stock Exchange (IDX) 2010-2014.
Wang (2011) explored the relationship between intellectual capital and financial performance using the data about companies from the capital market in Taiwan. He found a positive relationship between value-added intellectual capital indicator and return on assets in both fixed and random effects.
Ranani and Bijani (2014) investigated the influence of intellectual capital on the financial performance of listed companies on Tehran Stock Exchange. They found that it has a significant effect on the rate of return on assets. These findings were also confirmed by Arianpoor (2021), who noted that unrecorded and recorded intangible assets positively affect firm profitability (return on assets, profit margin). Finally, Seo and Kim (2020) analyzed Korean enterprises and found out that three intangible resources (human capital, advertising, and R&D) have a positive effect on a firm's profitability.

DATA AND METHODOLOGY
To test the assumptions highlighted in the paper, the activities of 16 Slovak companies' software developers were analyzed (Appendix A) for the period 2015-2019. These are innovative high-tech companies whose activities largely depend on the efficient use of intellectual capital.   To analyze the impact of relational capital on profitability, total costs of economic activity was chosen as an indicator. The last one is relevant because intangible assets do not include trademarks, as the capitalization of costs for their creation is prohibited by accounting legislation, as well as there are no other objects of accounting that characterize relational capital.

Profitability indicators References
However, according to Slovak legislation, costs for training and seminars, marketing and similar studies, market research, consulting, expertise, acquisition of standards and certifications such as ISO standards, preparation and commissioning, advertising, product marketing, restructuring and reorganization of the business, as well as other costs of a similar nature, are not accounted for as intangible fixed assets, but are accounted for as expenses for economic activities (Ministerstvo financií Slovenskej republiky, 2002). Therefore, the total costs of economic activity reflect the uncapitalized efforts of the enterprise aimed at increasing the relative capital of software development companies.
To control for a significant effect of additional variables except for main independent variables (intellectual capital indicators), this study gathered data on the company's size. Based on Ievdokymov et al. (2020b), company size was calculated as the natural log of total assets to control economies of scale.
For examining the relationships between the indicators that characterize the profitability of Slovak software development companies (ROA, EBITM, NPM, GPM -dependent variables), for indicators that characterize their intellectual capital (independent variables), and for control variables the paper must calculate the values of such indicators from the financial reporting. To fulfill this purpose and taking into account the features of accounting and financial reporting regulations in the Slovak Republic, a set of indicators and formulas for their calculation were used (Appendix B).
To investigate the relationship between intellectual capital and profitability indicators, this study examined the following models: Model 1 where ROA, NPM, GPM, EBITM -dependent variables, where i = entity and t = time; α -Identifier; β -Regression coefficient; CDC, S, AIFA, PC, SDC, SC, TEA, l_CS -independent variables, where i = entity and t = time; ε it -error term. Figure 1 shows the conceptual framework of the study.

Dynamics of profitability indicators of Slovak software developers
The development of four indicators that characterize the profitability of software developers in the Slovak Republic (ROA, NPM, GPM, EBITM) for the period 2015-2019 can be seen in Figure 2. To calculate the impact of intellectual capital on the profitability of Slovak software developers, the study uses information from the financial statements of intangible assets of such enterprises, as well as the costs Source: Authors' elaboration.

Descriptive test result
The descriptive statistics extracted from 16 Slovak software development companies from 2015 to 2019 provided a total of 80 observations, as tabulated in Table 2.
For all variables (dependent and independent) used to build the four proposed models, the basic numerical characteristics (mean, median, standard deviation, minimum, maximum) were calculated. The analysis of such characteristics allowed one to establish the following defining trends: for individual variables (ROA, GPM, EBITM, S, SDC, TEA, and l_CS) the mean value is greater than the standard deviation; thus, the data in these variables have small distribution. For all independent variables, except the control variable, there are significant gaps between the minimum and maximum values. As a result of their comparison with the corresponding median value, the paper encourages the use of robust statistics.

Classical assumption test
To confirm the compliance of the constructed FEM data, determine its adequacy, as well as to find ways to further improve the constructed model, it is necessary to diagnose it using a series of tests. Regression testing of panel data based on the Normality test, Autocorrelation test, and Heteroscedasticity test allows this paper to verify theoretical assumptions.
The results of testing the normal distribution of residues of all four models via Gretl software package revealed that the errors in them are abnormally distributed. For example, in Model 1, the value of Chi-square (2) = 11.896, p-value = 0.00261101, which is less than 0.05. It does not confirm the null hypothesis of a normal distribution of residues.
Autocorrelation testing should be performed to verify the relationship between one residue and another. To do this, the Wooldridge test was used for autocorrelation in panel data. It confirmed the presence of first-order autocorrelation for Models 1 and 3, and autocorrelation absence for Models 2 and 4. For example, for Model 2, the obtained p-value is higher than 0.05, confirming the null hypothesis of no first-order autocorrelation.
To determine the fact that there is an inequality of variants of the remains of one observation to another in the regression model, it must be analyzed for heteroskedasticity. Its absence is evidence of its better adequacy to the data used. To determine the level of heteroskedasticity for all four models, a nonparametric Wald test was used to establish its presence for all four models.
To improve the suggested models, it is also proposed to use robust estimators. It will prevent autocorrelation and correlation between units and variance problems in panel data regression models, detecting emissions, reducing their impact, or ex-cluding them from the statistical sample altogether. This practice is now widely used, particularly among scientists engaged in research in determining the role of intellectual capital in ensuring the profitability of enterprises (Dogan & Kevser, 2020).

Model 1 (ROA)
Tables 3-10 show the analysis performed using the FEM. It demonstrates the extent to which the independent variable will affect the dependent variable, given the non-measurable individual differences of objects, allowing one to get rid of the influence of an unobserved variable (constant in time) and get unbiased parameter estimates.  Note: * significant at the 10 percent level; ** significant at the 5 percent level; *** significant at the 1 percent level. where: ŷ -ROA, x 1 -CDC, x 2 -S, x 3 -AIFA, x 4 -PC, x 5 -SDC, x 6 -SC, x 7 -TEA, x 8 -l_CS.
In addition to the variables AIFA, SC, and TEA, all other selected independent variables for Model 1 are statistically significant. The most significant parameters are Const (P-value = < 0.0001), S (P-value = < 0.0001), PC (P-value = 0.0013), SDC (P-value = 0.0050), i.e., these indicators are the most influential on the dependent variable (ROA). The absence of corresponding asterisks in Table 3 indicates the insignificance of the intellectual capital indicator.
Equation of the Model 1 shows that part of the proposed parameters (const, AIFA, TEA, and PC) has a direct effect on ROA, and other parameters (CDC, S, SDC, SC, and l_CS) have an inverse effect. Table 4 indicates that the coefficient of determination (LSDV R-squared) of Model 1 is 0.33. This means that 33.5% of the variation of the dependent variable (ROA) can be explained by the variation of the selected independent variables (const, CDC, S, AIFA, PC, SDC, SC, TEA, and l_CS). At the same time, the rest is the impact of other variables that are not examined in this study.

Model 2 (NPM)
Model 2 can be demonstrated through the following equation:  Table 6 indicates that the LSDV R-squared of Model 2 is 0.39. This means that 39.1% of the variation of the NPM can be explained by the variation of the selected independent variables. Note: * significant at the 10 percent level; ** significant at the 5 percent level; *** significant at the 1 percent level.

Model 3 (GPM)
Model 3 can be demonstrated through the following equation: where: ŷ -GPM; x 1 -x 8the same as in Model 1.
In Model 3, only three parameters are statistically significant -CDC, PC, and l_CS ( Table 7). The most significant parameter is CDC (P-value = 0.0018), i.e., this indicator is the most influential on the dependent variable (GPM). Equation Model 3 shows that most independent variables (CDC, S, AIFA, PC, SDC, TEA, and l_CS) have a direct effect, and only two variables (const and SC) have an inverse effect on GPM. Table 8 indicates that the LSDV R-squared of Model 3 is 0.67, which is quite high compared to other models considered but not quite enough to talk about the significant role of intellectual capital in ensuring the profitability of enterprises. This means that 67.29% of the variation of the GPM can be explained by the variation of the selected independent variables.
In Model 4, only two parameters are statistically significant -CDC and l_CS ( Table 9). The most significant parameter, as in Model 3, is the parameter CDC (P-value = < 0.0001), i.e., this indicator is the most influential on the dependent variable (EBITM). Equation of the Model 4 shows that most independent variables (CDC, S, AIFA, SDC, SC, TEA, l_CS) have a direct effect, and only two variables (const, PC) have an inverse effect on EBITM.  Note: * significant at the 10 percent level; ** significant at the 5 percent level; *** significant at the 1 percent level.

CONCLUSION AND STUDY IMPLICATIONS
Obtained results partially confirm the previous studies on the positive impact of intellectual capital on the profitability of enterprises, as well as clarify which of its elements have the greatest impact on profitability indicators. Thus, the only independent variable that had a direct positive impact on the profitability of software development companies in all four analyzed models is Acquired Intangible Fixed Assets.
This study has some limitations. Firstly, future research aimed at confirming the hypothesis should cover not only the activities of software developers, as representatives of innovative enterprises, but also other economic entities belonging to this group of enterprises. Applying a more integrated approach will provide a comprehensive understanding of the role of intellectual capital in ensuring the profitability of innovative enterprises, as well as a comparison between enterprises in different industries, which will determine how to improve the policy of investing in intellectual capital depending on the industry.
Secondly, the study depends on the published financial data of Slovak software developers, so it is subject to all limitations that are inherent in the published financial statements of such companies. In addition, some companies that were not included in the study object either did not disclose financial statements for specific periods or had a limited period of existence (recently established or liquidated during the period under study). Therefore, they did not enter the panel data on which regression models were built.
The main theoretical assumption of this study that profitability of innovative enterprises depends to a small extent on the available intellectual capital, which is reflected in the financial statements, was justified by testing four panel data regression models.
The obtained results empirically confirm the conclusions of scientists on the unsuitability of conventional information provision in the treatment of intangible values and the need to expand the criteria Note: * significant at the 10 percent level; ** significant at the 5 percent level; *** significant at the 1 percent level. for recognizing intangible assets. Therefore, to increase the relevance of information on the intellectual capital of high-tech enterprises, it is recommended to expand the structure of financial reporting. Moreover, it is vital to improve the method of recognizing costs of various types (research and development costs, advertising and marketing costs, personnel costs, etc.) as intangible assets to increase management information requests.
The implementation of such proposals will help address the following important issues in the field of intellectual capital management: 1) Clearer identification of intellectual capital components in the financial statements of enterprises and ensuring unimpeded access to such information by different stakeholders; 2) Building a more effective system of internal control of intellectual capital, based on its new expanded structure; 3) Construction of a system of cost-oriented management of intellectual capital, based on an improved system of its accounting evaluation; 4) Development of a system for evaluating the effectiveness of the enterprise based on not only the material and financial, but also the intellectual capital; 5) Transformation of the management system of intellectual capital from tactical to strategic, based on the ability to obtain accounting information about the internal and external components of the intellectual capital in fair estimates.

ROA Return on Assets
ROA shows how profitable a company is in relation to its total assets Net Profit ROA Total Assets = Net turnover item (V1), Total assets item (S1) NPM Net Profit Margin NPM shows the extent to which the company earns money, i.e., how many cents of net profit was received for each euro of sale

Net Profit NPM
Total Sales = Income for the accounting period after tax (S100), Sales of goods (V3), Sales of own products (V4), Sales of services (V5), Sales of intangible fixed assets, tangible fixed assets and materials (V8) GPM Gross Profit Margin NPM shows the extent to which the company earns money, i.e. how many cents of gross profit (profit before subtraction selling, general, and administrative costs) was received for each euro of sale Note: * indicates the elements of the financial statements from which the data for the calculation of dependent and independent variables of four models were taken.