Parallels and Differences in Earnings Management of the Visegrad Four and the Baltics

Earnings management is a legal and widely preferred phenomenon of business finance that financial managers use to maintain and improve the enterprise’s competitiveness. Managers purposely manipulate business earnings to achieve the required status of the enterprise. The consequence of these activities is to provide a positive perspective for the owners, encourage the profitability for the creditor and the investors as well as demonstrate economic strengths to competitors. This article aims to identify parallels and differences in earnings management of enterprises in the Visegrad Four and the Baltics in terms of competitiveness for the nine-year period 2010-2018. The research uses a final sample of 4,543 observations from the EBITs of Slovak, Czech, Hungarian and Polish enterprises as well as 1,633 observations from the EBITs of Latvian, Lithuanian and Estonian enterprises. Time-series methods with all necessary assumptions have been run for the analyzed financial dataset. The results of the econometric modeling of unit roots show significant parallels in these groups of countries. The enterprises from the Visegrad group and the Baltics group use the apparatus of earnings management to be competitive. The obtained results confirm the systematic but legal manipulation from the side of management. A quantitative analysis of homogeneity tests using 1,000,000 Monte Carlo simulations indicates significant time differences of manipulation in these emerging countries. The year 2014 signaled a radical “accelerando” in earnings management for the V4, and the year 2016 is highlighted for the Baltics. fulfill the set goals. The apparatus of the Dixon test is used, followed by the Ljung-Box and McLeod-Li tests; Stationary Augmented Dickey-Fuller tests have also been conducted, supported by the Phillips–Perron test (No intercept, Intercept, Intercept and Trend considered), Homogeneity tests (von Neumann test and Standard Normal Homogeneity test), with their adequacy allowed by the Anderson–Darling test. Chapter Results describes the output of the examinations and the determinations regarding the hypotheses. This section identifies the parallels and differences in earnings management as a tool of competitiveness of enterprises within the V4 (Visegrad Four) and Baltic countries. In the Discussion part of the paper, current researches and investigations from developed countries are compared. Finally, brief conclusions are described, also presenting the limitations of our research and tracking possible tracks of future research. research has aimed to detect the use of earnings management in terms of competitiveness and to identify parallels and differences in earnings management of enterprises in the Visegrad Four and the Baltics. The outcomes of quantitative modeling have detected significant parallels in the analyzed countries regardless of the region. The development of earnings before interest and taxes of the Slovak, Czech, Polish, Hungarian, Latvian, Lithuanian and Estonian enterprises were not stationary during the nine-year period 2010-2018, which indicates the significant occurrence of earnings management. The time series analysis identifies not a homogenous development, but shows a two-year difference depending on the region. The radical acceleration of earning manipulation is marked in the year 2014 for the Visegrad group and for the Baltics group the year 2016 is the crucial one. The consequences of these results may be used not only in scientific fields but also in practice, i.e. for investors to determine more highly competitive


INTRODUCTION
Enterprises compete with each other only by the optimization of their own capital structures on the global market (Vagner, 2016). On the other hand, managers of enterprises doing business for more than a decade have a strong motivation to use aggressive behavior against their competitors Durana (Kljucnikov et al., 2016). Hudakova et al. (2018) support these conclusions by working with world surveys and studies which confirm that these strategies fundamentally contribute to improving the competitiveness of the enterprises. Management is facing challenges and requirements to the business model transformation (Malkawi & Khayrullina, 2021). Hoang & Joseph (2019) argue that earnings management represents a transformation towards a new accounting regime for many units. Financial managers generally prefer aggressiveness and legal earnings manipulation as indicators of sustainable manufacturing in Industry 4.0 (Susanto et al., 2019;Hayhoe et al., 2019). The use of risk management strategies to evaluate the routine activities of management assists enterprises in avoiding the pitfalls and management habits narrowly linked to a given rank of appropriate risk (Dvorsky et al., 2019). Kramarova & Valaskova (2020) highlight a conjunctive denominator: the use of information discrepancy within business finance to gain benefits. Meyers et al. (2019) emphasize the importance of earnings management strategies for enhancing decision-making using research results from big data. Thus, earnings management has become a very significant and especially trending issue for researchers nowadays. This practice can be considered routine in developed markets. However, in emerging countries it is still not very typical, although the economic factors are simpler to calculate from a methodological point of view (Belas et al., 2018). Our research uses a conventional earnings model to quantify the manipulation. We prefer original analysis focusing on the quantitative methods of time series, which is a new methodological approach using stationarity and homogeneity to indicate the manipulation of earnings. The main aim of our study is to identify earnings management strategies as well as parallels and differences in earnings management of enterprises in the Visegrad Four and the Baltics in terms of competitiveness for the nine-year period (2010-2018).
The article is divided as follows. The purpose of the research and the importance of earnings management in the context of competitiveness are depicted in the introduction. Subsequently, a literature review is provided focusing on analyses of various research studies that have represented the issue discussed. The third chapter shows our research objectives and hypotheses. Sophisticated quantitative methods using time series analysis have been chosen to fulfill the set goals. The apparatus of the Dixon test is used, followed by the Ljung-Box and McLeod-Li tests; Stationary Augmented Dickey-Fuller tests have also been conducted, supported by the Phillips-Perron test (No intercept, Intercept, Intercept and Trend considered), Homogeneity tests (von Neumann test and Standard Normal Homogeneity test), with their adequacy allowed by the Anderson-Darling test. Chapter Results describes the output of the examinations and the determinations regarding the hypotheses. This section identifies the parallels and differences in earnings management as a tool of competitiveness of enterprises within the V4 (Visegrad Four) and Baltic countries. In the Discussion part of the paper, current researches and investigations from developed countries are compared. Finally, brief conclusions are described, also presenting the limitations of our research and tracking possible tracks of future research.

THEORETICAL BACKGROUND
Earnings management has become a primary theme in accounting studies and a significant tool of researchers to determine instruments and mechanisms for making financial decisions regarding business finance and competitiveness (Sosnowski, 2018;Hoang, & Phung, 2019).
Recent studies on management earnings have focused on the several areas. Siekelova (2021) clarifies the historical development of the EM (earnings management) model and explains the development of profit models. Sosnowski & Wawryszuk-Misztal (2019a) show that some characteristics of management and supervisory board decisions improve the usefulness and credibility of earnings forecasts. Specifically, the more diversified the board in terms of age and the higher the management ownership results, the more accurate are the forecasts. Sosnowski & Wawryszuk-Misztal (2019b) add that certain characteristics of the management and supervisory board decisions can enhance the usefulness of forward-looking financial information from an IPO (initial public offering) prospectus, as some of characteristics of boards are important for the evaluation of the earnings forecast credibility at the time of IPO. The empirical results of Huynh (2020) show that the corporate governance mechanism is a significant moderation in the positive link between corporate social responsibility and earnings management. find that the pay disparities among the top management team (TMT) have a negative impact on real earnings management, and this relationship further enhances the value of enterprises which support the internal governance view. Lel (2019) deals with the role of international donors in restraining the earnings manipulation of enterprises; this study finds that managers manipulate the earnings less when independent foreign institutional investors are among their shareholders.  show that benchmark-beating enterprises entering into real earnings manipulation are less willing to engage in aggressive tax planning.  add that institutional investors holding stable equity stakes play an important monitoring role in reducing real earnings management by managers, who are pressured by capital market forces to "meet or beat" earnings targets. The research of Kundelis & Legenzova (2019) reveals that Estonian, Latvian and Lithuanian tax differences among the multinational corporations of parent and subsidiary countries may influence the subsidiary's profitable activity. A change of one percentage point among Estonia, Latvia, Lithuania and the foreign countries causes a 2.3% increase in a profit-shifting manner. Cugova & Cug (2020) comprehensively map the motivation for using earnings management techniques and models in the conditions of the current globalized market. Dyreng et al. (2020) provide a study of accrual and real earnings management in which the interests of shareholders and managers conflict. Their findings show that shareholders prefer avoiding violations by purposeful manipulation than violating a covenant without earnings management, with the former possibility ensuring shareholders a higher profit than the second. Khuong et al. (2020) have determined a positive link between real activities management and cash holdings detection, i.e. relevant diminutions in discretionary production costs and selling expenses enable managers to hide the real level of their financial situation of the enterprise. On the other hand, the negative correlation between accrual-based earnings manipulation and cash holdings shows that accruals support the moderation of the information asymmetry between the enterprise and other stakeholders. Valaskova & Fedorko (2021) monitor the manipulation with earnings in the transporting and storage sector (enterprises tend to manage earnings upwards) using the Beneish M-score.

RESEARCH OBJECTIVE, METHODOLOGY AND DATA
The paper aims to detect the use of earnings management in the competition and identify parallels and differences in earnings management of enterprises in the Visegrad Four and the Baltics for the nine-year period 2010-2018. The article focuses on cognate emerging markets. These countries are selected within the same investigation because earnings management has not been investigated and compared in these politically and economically close countries. Based on the historical development, countries of the V4 may be added to the Soviet-controlled Eastern bloc countries and countries of the Baltics to the formerly part of the Soviet Union. Joining the European Union in 2004 marks the next political parallel. The economic consistency of selected countries highlights the ranks of the Global Competitiveness Index (Table 1). The analysis allows formulating two hypotheses of parallels and differences in approaches to the competitiveness of enterprises from the Visegrad Four and the Baltics: H0A: There is a significant occurrence of earnings management practices in the countries of Visegrad Four and the Baltics.
H1A: There is no significant occurrence of earnings management practices in the countries of Visegrad Four and the Baltics.
H0B: There is no significant time difference in earnings management practices in the countries of Visegrad Four and the Baltics.
H1B: There is a significant time difference in earnings management practices in the countries of Visegrad Four and the Baltics.
The following statistical methods were implemented to gain the target of the paper and to test set hypotheses. They are described with their advantages and disadvantages. The significance level alpha 0.05 was used for all statistical tests.

The Dixon test to detect and remove outlying values.
Different approaches and ways are recommended to identify the occurrence of outliers and subsequently to process initial samples.

The Ljung-Box test and McLeod-Li test to reject serial correlation.
In financial time series investigations, it is required to check the serial correlations. The rejection of serial correlation confirms the fact that the used EBIT of analyzed enterprises from the Visegrad Four and the Baltics are independently distributed. Wong & Ling (2005) highlight that the conservative nature of the origin Box-Pierce statistic for testing serial correlation was improved. That is why the improved portmanteau statistics represented by the Ljung-Box test and McLeod-Li test are used. However, the size and the power performance of these two tests are not robust to heavily tailed data (Chen, 2002).

The Augmented Dickey-Fuller test and Phillips-Perron tests to identify the unit root.
A time series is stationary if its statistical properties do not vary in the process of time. A stationary time series marks that the mean and variance are constant over time. There are two possible approaches to test stationarity: unit root or stationarity test ∆y t = a 0 + γy t-1 + ε t (2) ∆y t = a 0 + γy t-1 +a 2 t + ε t The model applied for this augmented test: where ∆y t first-order linear differential of equation, γ unit root, The results are supported by the Phillips-Perron test with the same null hypothesis and the equation is estimated based on equations 1 to 3 depending on including an intercept, intercept and a trend, or neither a constant nor a trend. Test statistics: where γ is the estimator of γ, t γ is the t-fraction of (t γ =(γ/se(γ ))), se(γ) is the standard error of the regression, γ 0 is a consistent estimator of the variance error in the Dickey-Fuller regression, and it is obtained as (T-k) s 2 /T where k is the number of regressors and T is the number of observations, σ ε is an estimator of the residue spectrum.

The Anderson-Darling test to prove the normality.
The normality is a crucial assumption for using the chosen Standard Normal Homogeneity test (Kang & Yusof, 2012). More than 40 tests exist to prove the normality (Dufour, 1998

5.The von Neumann test and Standard Normal Homogeneity test to identify the heterogeneity.
The heterogeneity in the analyzed time series of financial data means that there is a significant  Table 2. These levels of purposive sampling are set because of the robustness and reliability of the research results and to analyze the enterprises with the same attributes (Singh & Masuku, 2014). The proper way to achieve representativeness is to form the populations by a relevant variable, which may lead to lesser weights for outliers but must be based on reliable information (Ghosh & Vogt, 2012

RESULTS AND DISCUSSION
Based on the observations of 1,089 Slovak enterprises, 722 Czech enterprises, 766 Hungarian enterprises, 1,966 Polish enterprises, 483 Latvian enterprises, 501 Lithuanian enterprises, and 649 Estonian enterprises, the annual average of EBIT is individually computed for the analyzed period ( Figure 1). The first hypothesis of the paper was firstly tested by the stationary augmented Dickey-Fuller test. Then these results were supported by the findings of the Phillips-Perron test. All three ways were run: the possibility with no intercept, with an intercept, and with an intercept and the linear trend.

Fig. 1 -Earnings before interest and taxes
As the p-value is greater than the significance level alpha, one cannot reject the null hypothesis. This situation means that the development of EBIT in the analyzed countries is not stationary (statistical properties vary in the analyzed period). The managers used earnings management practices to become competitive, as shown in a significant unit root. If the p-value is lower than the significance level alpha, the activities of managers do not show the marks of purposive manipulation. Based on the p-values of the stationary augmented Dickey-Fuller test and the Phillips-Perron test, the hypothesis of the existence of unit root cannot be rejected for all seven countries (Table 6 and Table 7). It is necessary to test the normality of analyzed financial time series in order to avoid rejecting the null hypothesis of the following of the normal distribution. It is required to test this assumption before the running the homogeneity tests used to evaluate the second hypothesis of the article. The Anderson-Darling test was computed. P-values (Table 8) are greater than the significance level alpha, which means that the data of enterprises from the Visegrad Four and the Baltics follow the normal distribution. The second hypothesis of the paper was firstly tested by the von Neumann test, following which the results were enriched by the findings of the Standard Normal Homogeneity test.
As the p-value of the von Neumann test is lower than the significance level alpha 0.05, the null hypothesis can be rejected, and the alternative hypothesis of the heterogeneity of time series is thus accepted. This situation means that the development of EBIT in analyzed countries is not homogenous. Thus a significant date was found in which managers were found to use earnings management practices significantly more intensively. If the p-value is lower than the significance level alpha, the activities of managers are the same, with no change in the frequency of manipulation. If a different intensity of the manipulation was shown, the time series of development of EBIT was divided into two periods. The Standard Normal Homogeneity test was supported by the von Neumann test and the significant date of changing the manipulation intensity was set. Based on the p-values of the von Neumann test and the Standard Normal Homogeneity test using 1,000,000 Monte Carlo simulations, the hypothesis of homogeneity for all seven countries can be rejected (Table 9 and Table 10). In addition, the Standard Normal Homogeneity test detected 2014 as the significant year for the Visegrad Group and the year 2016 for the countries of the Baltics. A great difference was confirmed among enterprises from the Visegrad Four and the Baltics that there is a significant time difference in the use of the earnings management practices as a tool for being competitive. Our results are discussed in our studies, which cover the significant incentive of earnings management from the analyzed countries of the Visegrad Four and the Baltics. Kramarova & Valaskova (2015) argue that the competitiveness and connected improving interdependence of the domestic market can affect the behavior of any economic unit, especially in terms of financial performance. This study analyzes transfer pricing and the effort to minimize tax manipulation in

CONCLUSION
Our research has aimed to detect the use of earnings management in terms of competitiveness and to identify parallels and differences in earnings management of enterprises in the Visegrad Four and the Baltics. The outcomes of quantitative modeling have detected significant parallels in the analyzed countries regardless of the region. The development of earnings before interest and taxes of the Slovak, Czech, Polish, Hungarian, Latvian, Lithuanian and Estonian enterprises were not stationary during the nine-year period 2010-2018, which indicates the significant occurrence of earnings management. The time series analysis identifies not a homogenous development, but shows a two-year difference depending on the region. The radical acceleration of earning manipulation is marked in the year 2014 for the Visegrad group and for the Baltics group the year 2016 is the crucial one. The consequences of these results may be used not only in scientific fields but also in practice, i.e. for investors to determine more highly competitive countries based on using unconventional indicators. These conclusions may also be helpful for the analysis regarding merge or fusion among these countries. An enterprise will be more precise by choosing partners for cooperation in terms of earnings manipulation. For governments, this study shows that they should increase their attention regarding whether the manipulation is always legal and they can use the presented methodological apparatus as a new paradigm detecting earnings management instead of or supplementing the conventional methods.
One weakness of the research is the short, nine-year period of the average EBIT analyzed as well as the unequal sample sizes of the enterprises from the individual countries. Limitations are also represented by the fact the whole sectors were analyzed together, not as separate groups that could have revealed the sectors with a higher or lower level of earnings manipulation. Future investigations can be focused on specific sectors, the extension of the studies for all European Union, dividing the countries according to membership in the European Monetary Union as well as setting a comprehensive methodology based on unit roots, stationarity and homogeneity to detect the occurrence of earnings manipulation and the identify the date and behavior incentives of significant acceleration for this activity. This research was conducted before COVID 19 pandemic. Future research may be undertaken from the year 2019 to disclose the impact of the crisis on the earnings management of the enterprises in regions of the Visegrad Four and the Baltics.