Analytical View on the Sustainable Development of Tax and Customs Administration in the Context of Selected Groups of the Population of the Slovak Republic

Abstract

Participation in the tax system of the country and tax liability itself are two of the basic characteristics of the functioning of all economically active subjects belonging to national economies. The revenues from taxes represent a substantial part of the resources in each state’s budget, so it is not easy to achieve any form of joint action by all concerned in this area. Within the revenues obtained by taxes, customs, and levies, the issue of taxes is one of the most sensitive areas. Higher efficiency of tax collection is possible to achieve by implementing not only external tools (e.g., tax control, detecting tax fraud, etc.) but also internal ones (e.g., internal motivation of taxpayers). Therefore, it is important to improve the perception of the tax system among the population and thus to influence and create a responsible approach to the collection of taxes and levies. Due to the COVID-19 pandemic, some changes and restrictions have been implemented into the system of tax and customs collection in the Slovak Republic, but very unexpectedly and purposefully. It results in a more complicated and confused system of tax and customs administration. Therefore, our research efforts are aiming to identify and analyze the key significant factors that affect the perception of tax and customs administration in the Slovak Republic from the perspective of its inhabitants, to capture differences in perception of taxpayers. The novelty of this study lies in its approach to identifying and eliminating redundancies and opacities in tax collection processes performed by institutions of tax and customs administration systems in Slovakia, namely the approach from the taxpayer point of view, to promote voluntary compliance with tax duties. Due to the lack of such oriented studies, a survey performed in three regions of Slovakia using the research tool (the questionnaire) developed by authors is presented in this article. To solve this complex problem, a solving strategy based on exploratory analysis (KMO = 0.641) and later confirmatory factor analysis is proposed to extract four key common factors from the items of the research tool (with a total value of 80.623% of the explained variability). The reliability of the used research tool, pre-tested in the pilot study, is demonstrated by the achieved Cronbach coefficient value of 0.827%. In order to evaluate the obtained data, the respondents were divided into groups by gender, residence, and occupation, in accordance with the indications and outputs of the pilot study. Finally, experimentally obtained data were analyzed and compared. The significant differences in the respondents’ perceptions of the extracted factors between selected groups are confirmed by applied analysis, both between men and women and between respondents depending on the place of residence. The main differences are observed in the questionnaire items Q10 “Administrative complexity of the tax and customs agenda”, Q7 “Ease of tax collection”, Q16 “Labor costs for staff dealing with the tax and customs agenda”, and Q5 “Helpfulness and willingness of administrative workers”. Time costs for tax processing are considered critical.

1. Introduction

One of the basic features of the functioning of all entities economically active in national economies is their participation in the tax system of the state and their own tax liability. For the comprehensive context of the tax system as a whole, it is important to know the impacts of the Slovak Republic’s accession to the European Union, as well as the financial administration in its states. The issue of taxation is one of the “most sensitive” areas. Each Member State’s tax revenue represents a substantial part of the revenue side of the state budget, which is why it is not easy to achieve any form of joint action by all stakeholders in this area [1]. The Financial Administration (FA) is a state administration office in the area of taxes, fees, and customs whose main task is to ensure the collection of taxes and customs in order to secure revenues to the state budget of the Slovak Republic and the EU. The successful collection of taxes and customs, thanks to the elimination of fraud, has brought about 3.7 billion euros more into the state treasury in recent years. Financial administration is a modern, client-oriented institution with the aim of intensifying electronic communications.

The issue of the effective collection of tax revenues is today one of the most significant problems faced by individual states, as public levies are the primary source of budget revenues [2]. At the same time, almost every state is obliged, to a greater or lesser degree, to pursue an overriding fiscal goal through shrinking excessive general government deficits in a manner that does not hamper the growth outlook for the state and simultaneously supports business activities [3]. A lack of adequate budget revenues hinders, and at times prevents, the state from functioning effectively [4].

The creation of a highly professional public sector and professional customs administration is a key issue for the sustainable development of national economies while employees are key persons for customs administration. Attracting and retaining talented employees are two of the strategic issues of the customs authorities in order to ensure the public good [5]. Effective and efficient organization of public administration and reorganization are some of the goals of policymakers and practitioners. Petersone and Ketners examine the processes of reorganization in the State Revenue Service in Latvia as a tool for authorities to achieve strategic objectives in order to make proposals for the planning activities of structural reforms. The research is mainly based on the monographic descriptive method, as well as the methods of analysis and synthesis. The main results are related to the identification of procedures and approaches for reorganization and structural changes in tax and customs administration. It can be concluded that, for the organizational structure, which is modelled according to the processes linked with units, it is possible to make precise analyses showing how the reorganization process will change performance indicators and clarify important factors. This research identifies the need for internal regulation and the respective procedure for implementing structural changes [6].

Quarantine economic conditions put forward new demands to tax administration’s efficiency as a priority source of the EU’s financial resources. The relationship between fair, efficient, sustainable taxation and budget revenues’ adequacy to finance social and economic investment, innovation, and development needs is evident. Therefore, ensuring tax administration’s efficiency and effectiveness in any EU country is of theoretical and practical importance [7].

The study [8] aims to deepen the theoretical and applied provisions of the VAT administration in e-commerce operations in a global convergence of the national economic systems. The analysis of the state and forecast of VAT mobilization in the EU countries’ e-commerce operations are presented. The research’s specific purpose is to assess the VAT tax base’s future size in the context of changes in EU tax legislation. According to the scientific literature analysis, in the framework of forecasting the studied tax, it is advisable to use econometric forecasting methods based on time series models. As the result of the study, the forecast volumes of the VAT base of Eastern Europe and the amount of approximate future tax revenues are presented. The obtained results confirm the relationship between the growth of e-commerce and VAT tax revenues under such transactions; the study presents author’s vision of the VAT administration ecosystem for e-commerce transactions and the allocation of its functional entities. The directions on achieving the maximum volumes of VAT mobilization following the EU tax legislation requirements in European integration and globalization were substantiated, and the conclusions can be used in tax administration processes in the matters of forecasting other tax payments [8].

The current processes of globalization are characterized by complexity, inconsistency, uncertainty, and, often, the predominance of politics over common sense. Effective, highly professional reform of the tax authorities and customs structures is a convincing example, not so much of a technological breakthrough but of the prospects and harmony of the new, innovative thinking of country leaders, government representatives, businesses, and citizens. Krutikov [9] reveals the features and problems present in the process of implementing digital economy technologies that prevent the provision of a level of cooperation adequate for modern requirements between national and foreign customs structures, as well as representatives of the business community. Proposals have been made to implement promising structural, institutional, and technological innovative approaches to creating customs activities using the positive experience of successful transformation of the tax authorities, which allowed for building a harmonious interaction with taxpayers, taking into account their expectations and needs. The rapid developments in the field of information and communication technologies (ICT) mean that tax authorities have to address the question of how to keep up with the latest trends in society. Digitalization is essential for providing better services to taxpayers, big data processing, and risk identification. It is also important for more efficient and effective tax collection.

One of the most important aspects of the sustainable development of tax and customs administration is the tax advantages of VAT, which provide an incentive for foreign trade participants and support their export activities [10]. As the burden of VAT is borne by the final consumer, the key issue here is the interaction between the state (represented by its controlling authorities) and the business actors in the process of tax administration [1].

The international scientific community actively debates various models of combatting fraudulent actions of entities leading to VAT gaps [11,12]. In particular, a team of authors analyze the factors defining the parameters of the VAT gap in an economy and conclude that three key categories of actors—final consumers, sellers, and the tax administration—must ensure the good-faith fulfilment of their duties [13]. Noting the negative impact of VAT fraud, certain researchers conclude that this illegal activity creates some of the most dangerous consequences for an economy [14,15] and proclaim that such activity can be implicated in the direct or indirect financing of terrorism [16]. The most important instrument for combatting VAT fraud is the development of international cooperation [17] and use of modern technologies [18].

Strengthening tax compliance is a key element of effective resource mobilization. Tax authorities are developing new strategies to promote voluntary compliance and to manage situations related to non-compliance or strategies to ensure cost minimization. The customs union plays an important role in protecting the external borders of the European Union and ensuring the fair and transparent collection of customs and other relevant taxes, such as excise duty and value-added tax. It also guarantees the smooth functioning of international trade between EU countries.

VAT is characterized as a tax defining the added value at an aggregate level and for a certain period; it imposes a tax on the difference between the total turnover and total purchases from different businesses [19]. Research in the field of VAT application in recent years has been focused more on improving the efficiency of tax control [20], improving the quality and efficiency of tax administration [21], minimizing the risks of tax fraud, and developing methods of reducing the scope of shadow business in general [22].

The European Commission is assisting EU Member States in reforming their tax policies and revenue management to ensure fair, sustainable, and efficient tax collection. It also supports EU Member States in reforming their public finance management and spending policies in order to achieve an efficient allocation of funds. A number of studies indicate that transformations in the field of taxation, including international taxation, ensure a sustainable development trend [23,24]. Kowal [18] highlighted the demonstrated efficiencies of a tax system through the degree of resistance that system poses to tax evasion.

The motivation of our research efforts is to promote strategies of voluntary compliance for tax, custom, and levy duties. The novelty of this study lies in its approach from the clients’ point of view to the identification and elimination of redundancies and opacities in tax/customs/levies collection processes performed by institutions of tax and customs administration systems in Slovakia. It is important to achieve the efficient and successful collection of taxes and customs by strengthening the voluntary compliance of clients (taxpayers), not only through external tools (improving the efficiency of tax control and of tax administration, clarifying tax frauds), as was mentioned above. Due to the lack of such oriented studies, a survey performed in three regions of Slovakia, using the research tool (the questionnaire) developed by the authors, is presented in this article.

The academic contribution of this research study lies in the fact that it points to areas that receive little attention in the scientific literature: identification of problematic components of the tax/customs system in Slovakia from the clients’ point of view. The subsequent analysis (exploratory analysis, confirmatory factor analysis) of data obtained by empirical research also contributes to filling gaps in the field of given scientific interests that are related to strengthening the successful collection of taxes, customs, and levies.

To solve this complex problem, a solving strategy based on exploratory analysis (KMO = 0.641) and later confirmatory factor analysis is proposed to extract four key common factors from the items of the research tool (with a total value of 80.623% of the explained variability). The reliability of the used research tool, pre-tested in the pilot study, is demonstrated by the achieved Cronbach coefficient value of 0.827%. In order to evaluate the obtained data, the respondents were divided into groups by gender, residence, and occupation, in accordance with indications and outputs of the pilot study. Finally, experimentally obtained data are analyzed and compared. The significant differences in the respondents’ perceptions of the extracted factors between selected groups are confirmed by applied analysis, both between men and women and between respondents depending on the place of residence. The main differences are observed in the questionnaire items Q10 “Administrative complexity of the tax and customs agenda”, Q7 “Ease of tax collection”, Q16 “Labor costs for staff dealing with the tax and customs agenda”, and Q5 “Helpfulness and willingness of administrative workers”. The results indicate dissatisfaction with the overall tax and customs administration. In particular, labour costs and time costs for tax processing are considered critical; this is further detailed in the Results and Discussion Section. The performed analysis presented in this study is one of the partial research tasks on the path converging to the goals set within the project VEGA 1/0194/19 supporting this research.

2. Materials and Methods

2.1. Research Focus

In order to promote strategies of voluntary compliance of tax, customs and levy duties, our research effort focuses on identification of problematic components of the tax/customs/levies system in the Slovak Republic from the clients’ point of view (taxpayer). In the center of our attention is the identification and later elimination of redundancies and opacities in tax/customs/levies collection processes performed by institutions of tax and customs administration systems in Slovakia. The study aims to identify and analyze the main indicators (factors) that influence the attitudes of citizens of the Slovak Republic to the functioning of tax and customs administration in order to capture the differences in perception of taxpayers. For these purposes, a survey performed in three regions of Slovakia using the questionnaire (the research tool developed by authors) is presented in this article. Based on the evaluation of data (outputs of the empirical research), the main and partial goals are stated.

The main goals of the research study are to determine significant factors that affect the perception of tax and customs administration in the Slovak Republic from the perspective of its inhabitants (taxpayers)—as respondents during the performed research and to extract the key significant factors from all factors expressed by individual items of the implemented research tool. Based on the obtained extracted factor structure, another partial aim is stated: to develop a factor structure model and verify it in the second group of respondents. The objectives presented in this study are the partial objectives of the tasks solved within the research project VEGA 1/0194/19.

2.2. Research Tool and Research Sample

The empirical research has been carried out using the research tool—the questionnaire (

Table 1. Structure of the research tool.

Num. of Quest.		Questions
Q1	-	Age
Q2	-	Gender
Q3	-	Occupation
Q4	-	Place of residence
Q5	-	Helpfulness and willingness of administrative workers
Q6	-	Competencies of tax and customs administration employees
Q7	-	Ease of tax collection
Q8	-	Ease of customs collection
Q9	-	Ease of levies collection
Q10	-	Administrative complexity of the tax and customs agenda
Q11	-	Time savings in the administration of the the tax and customs agenda
Q12	-	Availability of tax and customs administrations
Q13	-	User comfort thanks to digitalization of tax, customs, and levies administrations
Q14	-	Transaction costs for bank fees
Q15	-	Postage costs
Q16	-	Labour costs for staff dealing with the tax and customs agenda

). First-hand survey data have been collected over a two-year period starting on 1 June 2019 and ending on 1 June 2021. To perform the research, the questionnaires were distributed to potential respondents in two forms: (1) an electronic questionnaire distributed by email contacts and (2) a questionnaire in printed form administered in person to the chosen self-governing regions (available at the tax and customs offices). Both forms of the questionnaire were equivalent and could be used by respondents participating in the survey. One of the aspects used to divide the respondents into groups (for further intended analysis) was their employment, so the following groups were defined: students, employees, and entrepreneurs. Due to the fact that authors of this study are members of academic staff at universities operating in various self-governing regions (Prešov, Košice, Bratislava and Banská Bystrica), it was possible to give access to the questionnaire to students in both forms—in the electronical form (to email addresses) and in the printed form at faculties (anonymously). The survey protocol was strictly followed: data collection was accompanied by back-checks and other validation procedures, consistent with academic best practice.

A pilot study preceded the research presented in this paper. The pilot study was carried out from 2016 to 2018 and was attended by 2873 respondents. The results obtained by evaluation of the pilot study provided the bases for selection of the individual investigated/observed groups and for research tool modification. Taking into account the real distribution of respondents within the individual observed groups, the even stratified random sampling was chosen as most appropriate for the research purpose. The research tool presented in this paper is the result of the long-term research activities of the authors whose origins date back to the great reform of tax and customs administration in the Slovak Republic in 2014 and 2015. As mentioned above, the developed research tool was modified in 2018, after the evaluation of the pilot study. The items that did not confirm a significant relationship with the extracted factors were removed. Up to six common factors were extracted in a pilot study using exploratory factor analysis: Factor 1—collection and availability of levies, Factor 2—competencies of tax and customs administration employees, Factor 3—digitalization of tax and customs administration, Factor 4—efficiency of tax and customs collection, Factor 5—labor costs, and Factor 6—administrative complexity of the the tax and customs agenda. We assumed that the gradual implementation of changes in tax and customs administration, especially the digitization of tax and customs administration, would eventually result in a change in the respondents’ views on the functioning of tax and customs administration, which has been confirmed.

Another reason for the implementation of this study was the limitations resulting from the pandemic situation of COVID-19, especially at the turn of 2020 and 2021, when the measures of the Government of the Slovak Republic delayed the payment of levies and taxes. As a result of these facts, the factor model has been reduced to four dimensions from the original six. However, it is necessary to mention another tax reform in the Slovak Republic, which, according to available public resources, is being prepared for 2022. It does not yet have the full support of all parties in the Slovak government, but if this happens, it will be necessary to carry out research in this area again after new tax reform implementations are put into practice. Another modification of the research tool will correspond to this.

The final research tool—a questionnaire, consisted of four survey items (Q1 to Q4) and twelve evaluation items (Q5 to Q16). Questionnaire items Q2, Q3, and Q4 were measured on a nominal scale, while questionnaire item Q1 was measured on ordinal scale. The evaluation items of the research tool (Q5–Q16) were measured using a five-point Likert scale and represent variables on an ordinal scale. The reliability of the used research tool, pre-tested in the pilot study, is demonstrated by the achieved Cronbach coefficient value of 0.827%. This value represents an acceptable value confirming the reliability of the used research tool.

The final sample consists of 8328 respondents. The sample contains both urban and rural taxpayers and is geographically representative—at least at the level of regions. Subsequently, by a stratified random sampling, 1500 respondents were randomly extracted from this final sample to form relatively homogeneous groups (homogeneous from a quantitative point of view) with respect to the survey items Q1 to Q4. The respondents’ answers from this first group have been used for the exploratory factor analysis to develop the factor structure model.

From the remaining 6828 respondents, a second group of 1500 respondents was randomly extracted through the implementation of stratified random sampling in such way as to create homogeneous groups in quantitative terms with respect to items Q1 to Q4. Applying confirmatory factor analysis, the respondents’ answers from the second group have been used to verify the factor structure model developed in the first step. The analysis of the research sample itself points to the fact that out of the total number of N = 1500 respondents in the first group, there were 751 men (50.067%) and 749 women (49.933%). In terms of age structure (Q1), the first group included 384 respondents aged 18–25 (25.600%), 369 respondents aged 26–35 years (24.600%), 359 respondents aged 36–45 years (23.933%), and 384 respondents aged 46–60 years (25.867%). In terms of occupation, the research sample consists of 360 respondents who are students (24.000%), 373 employees (24.867%), 383 entrepreneurs (25.533%), and 384 respondents without statement of their occupation (25.600%). The Prešov self-governing region is identified as a place of residence (Q4) by 374 respondents (24.933%), 389 respondents are from the Košice self-governing region (25.933%), 352 respondents are from the Banská Bystrica self-governing region, and 385 are from the Bratislava self-governing region (25.667%). In terms of the place of residence, both regions of eastern Slovakia, one region of central Slovakia and the Bratislava self-governing region were selected intentionally. This intention was due to the comparison of respondents’ attitudes to the researched issue.

Table 2. Description of the research sample in terms of age, gender, and occupation of respondents.

N = 1500	Summary Table for All Multiple Response Items
	Q1	Q2	Q3 Student	Q3 Employee	Q3 Entrepreneur	Q3 Does Not State	Row Totals
Count	18–25 years	man	41	42	45	57	185
Column Percent			47.67%	48.84%	43.69%	52.29%
Row Percent			22.16%	22.70%	24.32%	30.81%
Table Percent			10.68%	10.94%	11.72%	14.84%	48.18%
Count	18–25 years	woman	45	44	58	52	199
Column Percent			52.33%	51.16%	56.31%	47.71%
Row Percent			22.61%	22.11%	29.15%	26.13%
Table Percent			11.72%	11.46%	15.10%	13.54%	51.82%
Count	Total		86	86	103	109	384
Table Percent			22.40%	22.40%	26.82%	28.39%	100.00%
Count	26–35 years	man	49	57	48	41	195
Column Percent			52.69%	61.96%	50.00%	46.59%
Row Percent			25.13%	29.23%	24.62%	21.03%
Table Percent			13.28%	15.45%	13.01%	11.11%	52.85%
Count	26–35 years	woman	44	35	48	47	174
Column Percent			47.31%	38.04%	50.00%	53.41%
Row Percent			25.29%	20.11%	27.59%	27.01%
Table Percent			11.92%	9.49%	13.01%	12.74%	47.15%
Count	Total		93	92	96	88	369
Table Percent			25.20%	24.93%	26.02%	23.85%	100.00%
Count	36–45 years	man	47	47	47	42	183
Column Percent			55.95%	52.81%	47.96%	47.73%
Row Percent			25.68%	25.68%	25.68%	22.95%
Table Percent			13.09%	13.09%	13.09%	11.70%	50.97%
Count	36–45 years	woman	37	42	51	46	176
Column Percent			44.05%	47.19%	52.04%	52.27%
Row Percent			21.02%	23.86%	28.98%	26.14%
Table Percent			10.31%	11.70%	14.21%	12.81%	49.03%
Count	Total		84	89	98	88	359
Table Percent			23.40%	24.79%	27.30%	24.51%	100.00%
Count	46–60 years	man	45	51	38	54	188
Column Percent			46.39%	48.11%	44.19%	54.55%
Row Percent			23.94%	27.13%	20.21%	28.72%
Table Percent			11.60%	13.14%	9.79%	13.92%	48.45%
Count	46–60 years	woman	52	55	48	45	200
Column Percent			53.61%	51.89%	55.81%	45.45%
Row Percent			26.00%	27.50%	24.00%	22.50%
Table Percent			13.40%	14.18%	12.37%	11.60%	51.55%
Count	Total		97	106	86	99	388
Table Percent			25.00%	27.32%	22.16%	25.52%	100.00%

Q1—age, Q2—gender, Q3—occupation.

reports a more detailed description of the research sample in terms of age (Q1), gender (Q2), and occupation (Q3).

The respondents of the second research set, whose answers have been used to verify the created factor structure model, are characterized by a such quantitative division into groups (according to the respective survey items Q1–Q4), which is almost identical to the first research set. There is evident a deviation at a level of ±3%. We do not present the last, fourth survey item of the research tool in Table 2 because it primarily serves as a comparative variable. In order to evaluate the obtained data, we divided the respondents into groups by gender, residence, and occupation, in accordance with the indications and outputs of the pilot study.

Let us present a few notes on the criteria division of respondents into groups. Already in the first study, carried out in the years 2016–2018, significant gender differences in the perception of individual items of the questionnaire were found among the respondents. Despite the current promotion of gender equality, there are, naturally, differences in the perception of different aspects of life between men and women. This fact is also confirmed by various studies by academic staff in the field of management, economics, and marketing, e.g., at the Faculty of Management and Business, University of Prešov [25]. Because these differences proved to be significant, we used them in this study as well.

Finally, experimentally obtained data were analyzed and compared to provide a suitable explanation for the results. A comparison of respondents’ views between groups in terms of their employment (a group of students, employees, entrepreneurs, and respondents without statement about their employment) will be part of a separate study that we intend to publish in the near future. Given the scope of this article, we will no longer mention them in this study.

3. Results and Discussion

Statistical data processing was performed by using Statistica 13.5 and IBM SPSS Amos 26. Factor analysis is a multidimensional statistical method that focuses on creating new variables in order to reduce large amounts of data with as little loss of information as possible. It is based on the assumption that the dependences between the observed variables are the result of the action of a certain smaller number of immeasurable quantities standing in the background, which are referred to as common factors. These common factors are defined as a linear combination of the original quantities.

Factor analysis has as its main goal to know and use the structure of common factors, which are considered to be hidden causes of mutually correlated variables, while it wants them to best explain and clarify the observed dependencies as simply as possible. This means that in the final solution, each variable should correlate with a minimum number of factors, and, at the same time, the number of factors should be as small as possible. The assumptions of factor analysis are more conceptual than statistical. Abnormality, heteroscedasticity, and nonlinearity do not affect the results of the factor analysis so much that we have to pay close attention to them. Some degree of multicollinearity is even desirable because the goal of factor analysis is to identify an internally bound, i.e., correlated group of features. The basic indicators of the suitability of using factor analysis are KMO statistics and the Bartlett test of sphericity. The Kaiser–Mayer–Olkin statistic (KMO) is an index that is used to compare the size of experimental correlation coefficients against the size of partial correlation coefficients. If the sum of squares of the partial correlation coefficients between all pairs of features is smaller than the sum of squares of the pairwise correlation coefficients, the KMO rate of the statistic is close to 1. Small values of KMO statistics indicate that factor analysis of original features will not be a good approach because it is not possible to explain correlation between pairs of features using other features. In our case, based on the value of the Keizer–Mayer–Olkin statistic (0.641) and based on the Kaiser definition for the correlation rate, it can be said that the choice of factor analysis for the analysis of data obtained using the above research tool is justified (

Table 3. Values of the KMO statistic and Bartlett’s test of sphericity.

Indicator		Value
Kaiser–Meyer–Olkin Measure of Sampling Adequacy		0.641
Bartlett’s Test of Sphericity	Approx. Chi-Square	357.287
	Df	66
	Sig.	0.000

Approx. Chi-Square—approximate value of χ2 test, Df—number of degrees of freedom, Sig.—the achieved level of significance.

). Bartlett’s sphericity test is a statistical test of the correlation between the original features. It tests the null hypothesis H₀: “There is no correlation between the characters”, so the correlation matrix is a unit matrix. The achieved value of significance of the Bartlett test of sphericity p = 0.000 is less than the chosen level of significance α = 5%, and, thus, we can reject the null hypothesis expressing that the implementation of a sample correlation matrix with 12 considered variables (Q5 to Q16) is a unit matrix. Based on the above information, we can conclude that factor analysis is suitable for the analysis of the obtained data.

One of the basic tasks of factor analysis is to reduce the number of original features (characters). When reducing the number of usable factors, certain basic criteria are taken into account, one of which is to first obtain, for a large number of original characters, a combination of characters that clarifies as much of the variance as possible. Then, combinations that relate to smaller and smaller variance are sought. Based on the above criteria for selecting the number of factors, we extracted four factors from the original matrix of characters by the method of main components (

Table 4. Eigenvalues of factor model.

Component	Initial Eigenvalues			Extraction Sums of Squared Loadings			Rotation Sums of Squared Loadings
	Total	% of Var.	Cum.%	Total	% of Var.	Cum.%	Total	% of Var.	Cum.%
1	1.657	13.805	13.805	1.657	23.805	23.805	1.291	20.760	20.760
2	1.117	9.308	23.113	1.117	19.308	43.113	1.264	20.530	40.190
3	1.091	9.089	32.202	1.091	19.089	62.202	1.172	19.764	60.954
4	1.023	8.522	40.724	1.023	18.522	80.724	1.160	19.669	80.623
5	0.996	8.299	49.023
6	0.957	7.973	56.996
7	0.936	7.800	64.796
8	0.900	7.503	72.299
9	0.885	7.378	79.677
10	0.841	7.008	86.685
11	0.812	6.763	93.448
12	0.786	6.552	100.000

Extraction Method: Principal Component Analysis, % of Var.—% of Variance, Cum.%—Cumulative%.

), whose eigenvalue is greater than 1 based on Kaiser’s criterion of statistical significance. Cumulatively, these four factors explain 80.623% of the total variance, with the first factor accounting for 20.760% of the variance, the second factor accounting for 20.530% of the variance, the third common factor accounting for 19.764% of the variance, and the fourth common factor accounting for 19.669% of the total variance by the factor model Varimax.

The factor matrix (

Table 5. Table of rotated factor loads of the factor model.

Question	FACTOR
	Factor 1	Factor 2	Factor 3	Factor 4
Q14	−0.606
Q10	−0.543
Q8	0.450
Q13	0.422
Q15		−0.735
Q11		0.595
Q12		0.431
Q6			0.618
Q16			−0.569
Q9			0.476
Q7				0.688
Q5				0.617

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.

) contains the factor loads for each character and in each factor. The non-rotated factor matrix represents the best linear combination of the original characters, and the best in terms of including as many character variabilities as possible. However, if the non-rotated factor matrix does not provide a meaningful definition of common factors, its rotation is approached. Factor load is an explanation of the role of each original feature in defining a common factor. It is actually the correlation coefficient between each original feature and factor.

From Table 5, we can see that the first common factor (Factor 1) significantly correlates with the questionnaire items Q14, Q10, Q8, and Q13. The values of the factor loads reach the value −0.606 for item Q14 and 0.422 for item Q13. From the analysis of the first common factor (Factor 1) presented in Table 5, it is clear that 36.724% of the variability of item Q14, 29.485% of the variability of item Q10, 20.250% of the variability of item Q8, and 17.808% of the variability of item Q13 are explained by the first common factor. The second common factor (Factor 2) correlates with items Q15 (−0.735), Q11 (0.595), and Q12 (0.431) of the given research tool. As with the first common factor, we can say about the second common factor that 54.023% of the variability of item Q15, 35.403% of the variability of item Q11, and 18.576% of the variability of Q12 are explained by the second common factor. As we can see from Table 5, in connection to the third common factor (Factor 3) it can be stated that 38.194% of the variability of item Q6, 32.376% of the variability of item Q16, and 22.658% of the variability of Q9 are explained by the third common factor. The last common factor (Factor 4) correlates with only two items of the research tool—item Q7 (0.668) and item Q5 (0.617). As shown in Table 5, 47.334% of the variability of item Q7 and 38.069% of the variability of item Q5 are explained by the fourth common extracted factor. We have thus verified the assurance of the practical significance of all three extracted factors.

Confirmatory factor analysis (CFA) was implemented on a sample of 1500 respondents to verify the defined factor structure. Although the polychoric correlation method is preferred for ordinal variables, in the current study, we chose the maximum likelihood method. When choosing the CFA method, we based our research on the results [26,27], which accept the application of classical cut-off estimators by the ML method even in the case of ordinal variables if they have at least five response categories. In this case, they can be considered interval variables. The distribution of responses to all items of the research instrument corresponded to the normal distribution. The normality of respondents’ responses was tested by the Shapiro–Wilks test at a significance level of α = 0.05.

In the first step, based on the construction of the factor model (Table 5), we defined the hypothetical structure of the data—factors, manifest variables, and relationships between them. We used basic and recommended indicators to test the suitability of the verified model. The basic recommended evaluation indicators [28] and their real achieved values on the applied four-factor model are given in

Table 6. CFA Fit Indices for factor model.

Fit Indices Used	Perfect Fit Indices	Acceptable Fit Indices	CFA Results	References
χ2/df	0 ≤ χ2/df ≤ 2	2 ≤ χ2/df ≤ 3	1.023	[29]
GFI	0.95 ≤ GFI ≤ 1.00	0.90 ≤ GFI ≤ 0.95	0.995	[30,31,32]
AGFI	0.90 ≤ AGFI ≤ 1.00	0.85 ≤ AGFI ≤ 0.90	0.991
CFI	0.95 ≤ CFI ≤ 1.00	0.90 ≤ CFI ≤ 0.95	0.996	[33,34,35]
NFI	0.95 ≤ NFI ≤ 1.00	0.90 ≤ NFI ≤ 0.95	0.972
TLI	0.97 ≤ TLI ≤ 1.00	0.95 ≤ TLI ≤ 0.97	0.995
RMSEA	0.00 ≤ RMSEA ≤ 0.05	0.05 ≤ RMSEA ≤ 0.08	0.004	[32,36,37,38]
SRMR	0.00 ≤ SRMR ≤ 0.05	0.05 ≤ SRMR ≤ 0.10	0.019
p-value	p > 0.05		0.428

χ2—Chi-square, df—degrees of freedom, GFI—goodness of fit index, AGFI—adjusted goodness of fit index, CFI—comparative fit index, NFI—The Bentler–Bonett normed fit index, TLI—Tucker–Lewis index, RMSEA—root mean square error of approximation, SRMR—standardized root mean square residual, p—probability level.

. The model applied in the research part is shown in Figure 1. The model assumes that the 12 variables, which are presented by the research tool items Q5 to Q16, are explained using four common factors: Factor 1—difficulty of tax and customs administration, digitalization of tax, customs and levies administrations; Factor 2—availability of tax and customs administration, time savings in administration; Factor 3—competencies of tax and customs administration employees; and Factor 4—ease of taxes and customs collection.

From Table 6, it is clear that the recommended indexes evaluating the factor model (Figure 1) are acceptable and entitle us to state that the created hypothetical model presents a perfect degree of agreement with real data and is applicable in such a form.

The analysis of the four-factor model of the view of selected groups of the population of the Slovak Republic on the sustainable development of tax and customs administration for the entire research group (N = 1500) is presented in

Table 7. Regression weights and standardized errors of the basic factor model.

Relationship			Estimate	Std. Estimate	Std. Error	t-Statistic	p-Value
Q14	←	F1	1.000	0.322	0.272	4.634	0.000 *
Q10	←	F1	0.852	0.243	0.198	4.315	0.000 *
Q8	←	F1	−0.924	−0.256	0.209	−4.427	0.000 *
Q13	←	F1	−0.873	−0.267	0.198	−4.408	0.000 *
Q15	←	F2	1.000	0.302	0.172	4.097	0.000 *
Q11	←	F2	−1.286	−0.369	0.284	−4.523	0.000 *
Q12	←	F2	−0.903	−0.301	0.208	−4.344	0.000 *
Q6	←	F3	1.000	0.208	0.297	3.887	0.000 *
Q16	←	F3	−0.957	−0.218	0.272	−3.521	0.000 *
Q9	←	F3	2.028	0.414	0.499	4.061	0.000 *
Q7	←	F4	1.000	0.349	0.313	3.411	0.000 *
Q5	←	F4	0.892	0.327	0.249	3.576	0.000 *

*—significant at the level of significance α = 0.05, Estimate—estimate, Std. Estimate—standardized regression weight, Std. Error—standard error, t-Statistic—value of t distribution, p-Value—probability level.

. From Table 7, it is clear that for the first common extracted factor (F1), concerning attitudes toward the complexity of tax and customs administration and toward digitalization, all exogenously directly measured variables are significant at the selected level of significance α = 0.05. The interrelationships between the exogenous variables, the items of the research tool in relation to the first common factor (F1), are positive for items Q10 and Q14. Based on the above, it can be said that the respondents agree with the claims that the system of tax and customs administration is administratively demanding, and they also perceive the transaction costs of bank fees this way. On the other hand, research tool items Q8 and Q13 are negative. It is therefore possible to conclude that the respondents do not consider the collection of duties to be simple and, at the same time, do not agree that the digitalization of tax and customs administration has increased user comfort. From the point of view of absolute value, they attach the greatest weight to the transaction costs of bank fees (0.322).

For the second common extracted factor (F2), in the sense of Figure 1, all bonds between endogenous and exogenous variables are also significant. It is interesting that we find negative links in the research tool items Q11 and Q12. Here, again, it can be concluded that respondents do not perceive time savings in the administration of the the tax and customs agenda and also do not consider tax and customs administrations to be available. For the third common extracted factor (F3), we observe a positive attitude towards the competence of tax and customs administration staff (Q6) and an equally positive attitude towards simplifying the collection of levies (Q9), but we also observe a negative attitude towards labor costs and customs agenda (Q16). In the case of the last, the fourth extracted factor (F4), we note the positive attitude of respondents to the willingness of employees of tax and customs offices (Q5) and, at the same time, a positive perception of the simplification of tax collection (Q7).

Table 8. Covariance and correlation of extracted exogenous variables.

Relationship			Covariance				Correlation
			Estimate	Std. Error	t-Statistic	p-Value	Estimate
F1	↔	F2	0.088	0.022	3.987	0.000 *	0.656
F1	↔	F3	−0.066	0.018	−3.601	0.000 *	−0.718
F1	↔	F4	−0.081	0.025	−3.270	0.001 *	−0.487
F2	↔	F3	−0.054	0.016	−3.324	0.000 *	−0.609
F2	↔	F4	−0.054	0.021	−2.61	0.009 *	−0.335
F3	↔	F4	0.068	0.021	3.252	0.001 *	0.615

*—significant at the level of significance α = 0.05, Estimate—standardized regression weight, Std. Error—standard error, t-Statistic—value of t distribution, p-Value—probability level.

shows the values of the coefficients of covariance and correlation between the extracted exogenous variables, i.e., factors. It is obvious that all mutual relations in terms of the factor model (Figure 1) are significant at the chosen level of significance α = 0.05. Based on the analysis presented in Table 8, we observe negative dependencies between the extracted factors F1 and F3 (−0.718), between the factors F1 and F4 (−0.487), between the factors F2 and F3 (−0.609), and between the factors F2 and F4 (−0.335). On the contrary, positive dependences on high levels are observed in the relationship between the extracted factors F1 and F3 (0.656) and the factors F3 and F4 (0.615). Additionally, the analysis of variance points to the fact that all factors are significant at the significance level α = 0.05, F1 (p = 0.000), F2 (p = 0.002), F3 (p = 0.015), and F4 (p = 0.005).

In the next analysis, we will focus on the basic differences in the perception of extracted factors according to the model shown in Figure 1, from the perspective of respondents’ gender and residence. The basic indicators of individual partial models, which define the correctness and applicability of individual models, are given in

Table 9. CFA fit indices for partial models.

Fit Indices Used	Gender		Residence
	M	F	PO	KE	BB	BA
χ2/df	0.932	0.894	1.013	0.739	1.012	0.880
GFI	0.991	0.991	0.981	0.986	0.979	0.984
AGFI	0.984	0.985	0.965	0.975	0.964	0.971
CFI	1.000	1.000	0.995	1.000	0.991	1.000
NFI	0.952	0.978	0.955	0.986	0.921	0.908
TLI	0.992	0.994	0.991	0.994	0.987	0.992
RMSEA	0.000	0.000	0.006	0.000	0.006	0.001
SRMR	0.025	0.024	0.036	0.031	0.038	0.034
p (χ2)	0.600	0.673	0.448	0.899	0.450	0.697

χ2—Chi-square, df—Degrees of freedom, GFI—goodness of fit index, AGFI—adjusted goodness of fit index, CFI—comparative fit index, NFI—The Bentler–Bonett normed fit index, TLI—Tucker–Lewis coefficient, RMSEA—root mean square error of approximation, SRMR—standardized root mean square residual, p—probability level, M—man, F—woman, PO—Prešov Self-governing region, KE—Košice Self-governing region, BB—Banská Bystrica Self-governing region, BA—Bratislava Self-governing region.

. Permissible values of individual indices are given in Table 6.

Table 10. Standardized regression weights and errors of partial models from the perspective of the gender of respondents.

Relationship			Man				Woman
			Std. Estim.	Std. Err.	t	p	Std. Estim.	Std. Err.	t	p
Q14	←	F1	0.429	0.820	2.261	0.007 *	0.266	0.206	3.114	0.001 *
Q10	←	F1	0.189	0.179	2.608	0.009 *	0.309	0.400	3.480	0.000 *
Q8	←	F1	−0.223	0.209	−2.892	0.004 *	−0.258	0.350	−3.238	0.001 *
Q13	←	F1	−0.273	0.216	−3.065	0.002 *	−0.243	0.325	−2.993	0.003 *
Q15	←	F2	0.260	0.191	2.996	0.003 *	0.294	2.480	−3.004	0.004 *
Q11	←	F2	−0.464	0.641	−2.857	0.004 *	−0.318	0.341	−3.440	0.000 *
Q12	←	F2	−0.183	0.260	−2.378	0.017 *	−0.409	0.367	−3.551	0.000 *
Q6	←	F3	0.124	0.659	2.111	0.047 *	0.301	0.403	−4.211	0.000 *
Q16	←	F3	−0.116	0.534	−1.686	0.092	−0.265	0.237	−3.234	0.001 *
Q9	←	F3	0.333	1.189	2.300	0.021 *	0.484	0.475	3.471	0.000 *
Q7	←	F4	0.497	0.311	2.278	0.006 *	0.258	0.297	3.269	0.014 *
Q5	←	F4	0.254	0.222	2.196	0.028 *	0.401	0.652	2.271	0.023 *

*—significant at the level of significance α = 0.05, Std. Estim.—standardized regression weight, Std. Err.—standard error, t—t-statistic, p—probability level.

contains estimates of standardized regression weights and their statistical significance for the categories of male and female respondents. The main differences can be seen in the perception of item Q16 of the research tool (labor costs for staff dealing with the tax and customs agenda). While women consider this item to be statistically significant in terms of perceptions of the functioning of tax and customs administration, men consider this item to be irrelevant. Men consider item Q7 (0.497) as the most important, and, therefore, for them, from the point of view of tax and customs administration, the most important is the simplicity of tax collection, followed by item Q11 (−0.464), which concerns time savings in tax and customs administration and, finally, item Q14 (0.429)—transaction costs for bank charges. For women, the most important item is Q9 (0.484), which concerns the area of simplicity of levies collection, followed by the research tool item Q12 (−0.409) focusing on the availability of tax and customs administration and, finally, item Q11 (−0.318), to which men attach the greatest importance. From the differences in point of view on the significance of individual items between men and women, we observe the largest difference in the perception of item Q7, where the value of the difference between men and women is 0.239. Men attach more importance to this item. The second largest difference is observed in the research tool item Q12 (0.226), where, again, men give it more weight. For item Q6, which concerns the competence of tax and customs administration staff, the difference is −0.177, which means that the competence of employees is more important for women than for men, and, at the same time, the competence of employees is perceived more positively by women. As in the previous case, Q9 is perceived more positively by women than men, with the value of the difference between the two groups being −0.151. Other significant differences in the perception of the importance of the individual items of the research instrument are observed between men and women in item Q16 (0.149), which concerns labor costs dealing with the the tax and customs agenda, with women perceiving this area more negatively than men. For item Q5, focusing on the perception of the willingness of tax and customs administration staff, the value of the difference is at the level of −0.147, with women attaching more importance to this item. The last significant difference in the respondents’ answers is in the item of research tool Q11 (−0.146), wherein this item is perceived more negatively by men than women.

The analysis of mutual relations between extracted factors is given in

Table 11. Analysis of mutual relations of factors of partial models from the gender point of view of respondents.

Relationship			Man			Woman
			Covar.	p-Value	Correlat.	Covar.	p-Value	Correlat.
F1	↔	F2	0.072	0.021 *	0.446	−0.140	0.000 *	−0.849
F1	↔	F3	−0.062	0.030 *	−0.844	0.067	0.005 *	0.590
F1	↔	F4	−0.138	0.003 *	−0.436	−0.120	0.004 *	−0.573
F2	↔	F3	−0.045	0.052	−0.964	−0.049	0.011 *	−0.452
F2	↔	F4	−0.062	0.048 *	−0.310	0.061	0.060	0.306
F3	↔	F4	0.065	0.036 *	0.712	−0.066	0.012 *	−0.478

*—significant at the level of significance α = 0.05, Covar.—Covariation, Correlat.—Correlation. p-Value—probability level.

. We can see that the fundamental difference between men and women is in the perception of the relationship between factors F2 and F3. This relationship is insignificant for men (p = 0.052) in contrast to women, who attach importance to this relationship (p = 0.011). We see the opposite trend in the relationship between factors F2 and F4. In contrast to the previous case, this relationship is insignificant for women (p = 0.060), but men attach importance to it (p = 0.048). Other significant differences between men and women are observed in the interrelationships between the F1 and F2 factors. While men attach medium importance to this relationship at the level of the correlation coefficient of 0.446, and the relationship is positive, i.e., if the first factor was perceived positively by the respondents, then the second extracted factor was also positively perceived. In women, with the same relationship, we observe a very strong (−0.849) but negative relationship. Thus, the tendency of women to perceive the common relationship of the F1 and F2 factors is the opposite of that of men. We also observe a conflicting perception of the relationship between factors between men and women in the correlation of factors F1 and F3. While men perceive this relationship of factors very negatively (−0.844), women perceive this relationship positively but at a medium level of strength (0.590). We also observe an ambivalent perception of the interrelationships between the extracted factors between men and women in the relationship between factors F3 and F4. In both comparison groups, this relationship is statistically significant, but men perceive it very positively (0.712) in contrast to women, who perceive this relationship negatively (−0.478).

The second part of the analysis deals with the description of differences in the perception of tax and customs administration between respondents depending on their place of residence. The aim is to identify differences in the perception of the observed issue in relation to the region where respondents live or work. Both regions of eastern Slovakia were selected in a targeted manner: the Prešov and Košice self-governing regions. Furthermore, one representative of Central Slovakia was selected: the Banská Bystrica self-governing region. Western Slovakia was represented by the Bratislava elf-governing region. We start from the assumption that the respondents of the Prešov and Košice self-governing regions will show a high degree of agreement in the perception of individual extracted factors and the relationships between them. On the other hand, we assume relatively large differences in the perception of tax and customs administration between the respondents of the Prešov and Košice self-governing regions and the respondents from the Bratislava self-soverning region. The answers of the respondents from the Banská Bystrica self-governing region should coincide with the answers in the perception of individual factors with the respondents from eastern Slovakia. Individual indicators of the suitability of partial models, which are based on the basic factor model presented in Figure 1, are defined in Table 6. Based on Table 6, it can be stated that all partial models of the perception of tax and customs administration based on the respondents’ residence place are correct. Estimates of standardized regression weights and their achieved level of significance are given in

Table 12. Standardized regression weights of partial models for respondents’ residence.

Relationship			PO		KE		BB		BA
			Std. Estimate	p-Value	Std. Estimate	p-Value	Std. Estimate	p-Value	Std. Estimate	p-Value
Q14	←	F1	0.386	0.042 *	0.404	0.000 *	0.321	0.002 *	0.375	0.005 *
Q10	←	F1	0.186	0.055	0.349	0.000 *	0.192	0.059	0.287	0.018 *
Q8	←	F1	−0.414	0.009 *	−0.324	0.000 *	−0.205	0.053	−0.022	0.792
Q13	←	F1	−0.229	0.028 *	−0.216	0.010 *	−0.341	0.016 *	−0.299	0.034 *
Q15	←	F2	0.437	0.002 *	0.323	0.002 *	0.247	0.043 *	0.305	0.041 *
Q11	←	F2	−0.326	0.006 *	−0.473	0.006 *	−0.476	0.021 *	−0.199	0.105
Q12	←	F2	−0.191	0.041 *	−0.318	0.015 *	−0.389	0.022 *	−0.223	0.091
Q6	←	F3	0.213	0.027 *	0.230	0.007 *	0.058	0.075	0.309	0.048 *
Q16	←	F3	−0.234	0.044 *	−0.354	0.003 *	−0.111	0.493	−0.079	0.375
Q9	←	F3	0.416	0.020 *	0.334	0.004 *	0.691	0.608	0.485	0.045 *
Q7	←	F4	0.349	0.013 *	0.303	0.000 *	0.219	0.039 *	0.387	0.036 *
Q5	←	F4	0.574	0.096	0.357	0.007 *	0.190	0.183	0.271	0.039 *

*—significant at the level of significance α = 0.05, Std. Estimate—standardized regression coefficient, p-Value—probability level, PO—Prešov self-governing region, KE—Košice self-governing region, BB—Banská Bystrica self-governing region, BA—Bratislava self-governing region.

.

The initial assumption that the opinions and attitudes of respondents from the Prešov and Košice self-governing regions will be similar can be rejected. As we can see from Table 12, the most significant difference is in item Q10 of the research tool, namely in the opinions of respondents on the administrative complexity of the tax and customs agenda. Respondents from the Košice self-governing region attach importance to item Q10 (p = 0.000), unlike to respondents from the Prešov self-governing region, for whom this item is not important (p = 0.055). The second significant difference can be seen in item Q5, namely the perception of the willingness of employees of tax and customs offices, where this willingness is not statistically significant for respondents from the Prešov self-governing region (p = 0.096) as opposed to respondents from the Košice self-governing region. Here it should be said that all items of the research tool and factor model (Figure 1) are significant at the level of significance α = 0.05 for respondents from the Košice self-governing region. In terms of the importance of significant items of the research tool, it can be stated that while item Q15 is the most important item for respondents from the Prešov self-governing region, for respondents from the Košice self-governing region, it is item Q11, i.e., time savings in the administration of the tax and customs agenda, which respondents perceive negatively (−0.473).

When asking about the assumption of a high level of consent of respondents from the Prešov and Košice self-governing regions, it is necessary to state several facts. In terms of industry and economy, there are differences in real life between the various regions of Slovakia. It is interesting to observe, for example, the distribution of major manufacturing companies. The industry is concentrated mainly in the area of western and central Slovakia, while there are only a few large employers in the area of eastern Slovakia compared to the other regions. A complementary phenomenon to the previous statement is the similarity of the unemployment rate and the average wage in the Prešov and Košice self-governing regions. There are also observable/evident large differences in other evaluated items between the regions of eastern Slovakia and the rest of Slovakia, of course to the detriment of eastern Slovakia. The obvious presence of these facts led us to assume that the attitudes of respondents from the Prešov and Košice self-governing regions would be similar

When examining the answers of respondents from the Banská Bystrica self-governing region, we can see that in comparison with the previous two groups, there are differences in the perception of the importance of individual items, in particular in the following items: Q8—expressing the simplicity of customs collection (p = 0.053) and Q6 (p = 0.075), dedicated to the competence of tax and customs administration staff. We also record differences in item Q16 (p = 0.493) focused on labor costs dealing with the tax and customs agenda and finally in item Q9 (p = 0.608), which deals with the simplicity of the levies collection. These listed items of the research tool are not significant for respondents from the Banská Bystrica region. The most important item of the research tool for respondents from the Banská Bystrica region is the time saving in the administration of the the tax and customs agenda (Q11) with the value of the standardized regression coefficient at the level of −0.467. As in the other groups mentioned above, we observe a negative perception of this item, but with the greatest weight for respondents from the Banská Bystrica self-governing region during all compared groups. We observe a certain similarity in the perception of tax and customs administration among respondents from the Banská Bystrica and Bratislava self-governing regions, in view of the simplicity of customs collection. It is declared by item Q8 which is statistically insignificant for both groups, unlike to respondents from the Prešov and Košice self-governing regions. In a similar sense, respondents also perceive item Q16, namely labor costs that deal with the tax and customs agenda. Respondents from the Prešov (p = 0.044) and Košice (p = 0.003) self-governing regions consider these costs to be important and significant in contrast to the respondents from the Banská Bystrica (p = 0.493) and Bratislava (p = 0.375) self-governing regions, who do not attach importance to them.

A certain degree of similarity in the perception of tax and customs administration among respondents from the Košice and Bratislava self-governing regions can be seen in item Q10, concerning the administrative complexity of tax and customs administration. These respondents consider the item to be significant at the level of significance α = 0.05 in contrast to the respondents from the Prešov and Banská Bystrica self-governing regions, who do not attach importance to it. The willingness of employees of tax and customs offices represented by item Q5 is important for respondents from the Bratislava (p = 0.039) and Košice (p = 0.007) self-governing regions in contrast to respondents from the Prešov (p = 0.096) and Banskobystrický (p = 0.096) self-governing regions, for whom this item is irrelevant. For the respondents from the Bratislava self-governing region, the most important item of the research tool is Q9 (0.485), the simplicity of the collection of levies. The second most important indicator of the perception of tax and customs administration by respondents from the Bratislava self-governing region is the simplicity of tax collection, which is represented by item Q7 (0.387). The third most important item is item Q14 (0.375), which concerns the transaction costs of bank fees. Within the mutual similarity of the perception of tax and customs administration by all groups of respondents, they hold equally significant positions in Q14 (transaction costs for bank fees), Q13 (user comfort due to digitalization), Q15 (postage costs), and Q7 (ease of tax collection). Other differences in the perception of respondents’ attitudes to the tax and customs administration are described in Table 12.

From the point of view of mutual relations between the extracted factors in the form of their mutual correlation, in

Table 13. Correlations between partial model factors for respondents’ residence.

Relationship			PO		KE		BB		BA
			Correl.	p-Value	Correl.	p-Value	Correl.	p-Value	Correl.	p-Value
F1	↔	F2	0.519	0.033 *	0.651	0.013 *	0.822	0.038 *	0.492	0.134
F1	↔	F3	−0.589	0.062	−0.951	0.005 *	−0.325	0.617	−0.434	0.095
F1	↔	F4	−0.288	0.173	−0.602	0.020 *	−0.061	0.883	−0.723	0.015 *
F2	↔	F3	−0.866	0.024 *	−0.589	0.035 *	−0.253	0.622	−0.601	0.089
F2	↔	F4	−0.026	0.870	−0.224	0.283	−0.391	0.288	−0.587	0.069
F3	↔	F4	0.454	0.151	0.937	0.015 *	0.638	0.617	0.451	0.101

*—significant at the level of significance α = 0.05, Correl.—value of correlation coefficient, p-Value—probability level, PO—Prešov self-governing region, KE—Košice self-governing region, BB—Banská Bystrica self-governing region, BA—Bratislava self-governing region.

, we see significant differences in their perception across individual groups of respondents. Respondents from the Prešov self-governing region consider the relationship between factors F1 and F2 (p = 0.033) to be important, with a medium strength of the mutual relationship with the value p = 0.519. The second significant relationship is observed between factors F2 and F3 (p = 0.024), but a strong but negative relationship is evident here. For the respondents from the Košice self-governing region, we observe all correlations in Table 13 as significant, except for the relationship between factors F2 and F4 (p = 0.283). The mutual significance of factors F1 and F3 (−0.951), followed by the relation of factors F3 and F4 (0.937) is of the greatest importance for the respondents.

Many developed countries have and continue to address the issue of uniform collection of taxes and levies. Some states have managed to translate the answers to the problematic questions into a functioning system, while others still remain in the well-established ways of separate tax collection from premium collection.

Increasing administrative costs, in conjunction with a system that has separate collection of taxes and levies, reduces the efficiency of their collection, thus significantly reducing the scope (area) for permanently reducing tax and levy rates or providing better, higher quality public services. We currently consider the selection system to be complicated, administratively and financially demanding not only for public finances, but also for those it is intended to serve, i.e., for citizens and businesses. Entrepreneurs, or more accurately, employers have to fulfill certain obligations in reporting and paying taxes and levies every month they are obliged (in most cases) to communicate with three institutions that, however, differ materially. These are the tax office, the social insurance agency, and the health insurance company, creating another complication in connection with health insurance because there is a need to communicate with a certain health insurance company depending on where the individual employees are insured.

The collection of taxes and individual levies takes place in three different institutions, which, as mentioned above, reduces the efficiency of the collection of taxes and levies and reduces the success of the recovery of individual receivables in the event of non-payment of taxes and levies. At the same time, such a system helps to utilize methods of debt collection and assignment. The increase in operating costs of individual institutions is a consequence of the lack of interconnection of their information systems. As a result of these facts, we observe a negative impact on the ever-increasing growth in bureaucracy and enlargement of the administrative apparatus, which, of course, also reflects in the negative direction the expenditure part of the country’s budget. Three ministries, three factually different institutions, and three different laws also cause different approaches to solving the same—annual settlement of income tax by tax authorities, annual settlement of health contributions by health insurance companies, non-settlement of social contributions by the social insurance agency, different assessment bases for social and health contributions, different databases, different information systems, and so on.

Based on the aforementioned factors, it is clearly more advantageous and contributes to the effectiveness of the PS (pension system) to implement a joint collection of taxes and levies through one institution. A much simpler system, which would allow the calculation of taxes and levies within a single form, would be an advantage for payers (citizens and businesses). In this case, all payments and information about them would go to only one institution, which would be the only partner of natural and legal persons. After the approval and implementation of the reform, there would be only one central database of citizens—tax entities and policyholders at the same time. This institution should also be the single point of contact in the event of collection problems, which will make it easier for taxpayers or, to be more accurate, premium payers, to access the private sector and reduce the complexity of complicated administrative processes for the public sector.

In the optimal case, it is possible to simultaneously implement the unification of the payment of taxes and social security and health insurance. The issue of the unification of health contributions would require additional analyses in view of the larger number of health insurance companies. However, the fact is that, despite the initial problems, a similar model for pension management companies is proving to be effective, even from the point of view of the pension management companies themselves.

There are two basic areas of concern that complicate the transition to a single collection of taxes, duties, and insurance levies. The first is the preparation of institutions that will collect taxes and insurance levies. However, the reform of the uniform collection of taxes, duties, and insurance levies is a relatively complicated process. A change in the method of collecting levies is considered in the first place to be a substantive decision which should ensure, as far as possible, the highest possible payment discipline at the lowest administrative costs. This optimization of the collection of taxes and levies is not related to any politically sensitive issues such as the number of levies or the design of the relevant benefits. The merging of the collection of levies will be a significantly effective step in properly ensuring the implementation and enforcement of the planned reform, regardless of the degree of solidarity or merit or how much these aspects will be incorporated into the calculation of benefits. The decision on the number of levies or benefits will also have no effect on this.

The primary expectation for the implementation of the reform of the unification of the collection of taxes and levies is the absolute reliability and quality of the financial administration information system, which should have the task of transferring funds and a database with information on payments of individual policyholders and employers to the social insurance agency. The individual accounts of taxpayers and policyholders must be prepared very carefully. The information will be extremely sensitive, as it will form the basis for the payment of benefits to the Social Insurance Agency. Already at the beginning of the implementation of the reform, it is necessary to clearly state that only the collection of insurance contributions, including the control and subsequent recovery of receivables, would pass to the financial administration, while the expenditure part, especially the calculation and payment of benefits, would remain within the competence of the social insurance agency.

The main difference in the perception of tax and customs administration between men and women is observed in the questionnaire item Q16 (labor costs for staff dealing with the tax and customs agenda), which women consider significant (p = 0.000) while men consider this item insignificant (p = 0.092). From the point of view of respondents’ gender, the most significant difference in the attitudes of respondents occurs in item Q7 (ease of tax collection) with the value of difference significance at the level of 0.239, wherein this item is more important for men (0.497) than for women (0.258). The most significant difference between the two self-governing regions of eastern Slovakia is in item Q10, “Administrative complexity of the tax and customs agenda.” Respondents from the Košice self-governing region attach the importance of Q10 (p = 0.000) in contrast to respondents from the Prešov self-governing region, for whom this item is not important (p = 0.055). The second significant difference can be observed for item Q5, “Helpfulness and willingness of administrative workers”, where this item is not statistically significant for respondents from the Prešov self-governing region (p = 0.096) as opposed to the respondents from the Košice self-governing region (p = 0.007).

The present study assesses the current situation. Thus, it was implemented 4 years after the last major reform of the tax and customs administration. The aim was to verify the respondents’ views on the functioning of the tax and customs administration 4 years apart and to find out the key significant factors. It is also necessary to mention the fact that within the Slovak Republic, there are frequent changes in the field of research. This causes changes in respondents’ views on the functioning of the tax and customs administration over time.

It would certainly be necessary and interesting to compare the results with other countries in the EU and especially in the V4 countries (Hungary, Poland, Czech Republic), which are closest to Slovakia both in terms of very similar historical development and in terms of economic development as post-communist countries. However, we do not have similar data from these countries because the research tool (the questionnaire) was created by the authors of the presented study. However, as part of the project solution, thanks to which we carry out the research, we plan to expand the research and compare with the Czech Republic. This project intention is proposed for the second half of 2022.

The limitations of this study are given primarily by the nature of the research focus. Due to the frequent changes in the laws that affect the tax and customs administration, the study is limited to evaluating the functioning of the tax and customs administration only through the subjective perception of the respondents, i.e., through their views and attitudes. As a result, we have reduced the number of items in the research tool and tried to simplify them as much as possible. Other limitations of this study were caused by unexpected events that occurred during the implementation of data collection in the period from 01 June 2019 to 01 June 2021. Due to the COVID-19 pandemic, changes and restrictions were unexpectedly and purposefully adopted, which greatly complicated the already complex system of tax and customs administration and made it even more confusing. The government of the Slovak Republic took steps that significantly worsened the perception of the functioning of the tax and customs administration, especially among entrepreneurs, which we observed in their answers during 2021. We mean such steps as the government extending/postponing the deadlines for levies and taxes. However, the tax authorities made entries in the list of debtors, imposed penalties, and executions. Other limits can also be perceived from the point of view of the geographical location of the respondents. In principle, more negative answers were from entrepreneurs in the east of Slovakia (the Prešov and Košice self-governing regions). However, due to changes in the course of the study (especially the turn of 2020 and 2021 and the whole period until 1 June 2021), the results are heterogeneous in time, which allowed us to capture the time change in the perception of the tax and customs administration. The next direction of research will be to capture the perception of tax and customs administration in the period of the global pandemic COVID-19 in a smaller sample of about 1100 respondents and, after the adoption and implementation of tax reform, to conduct new research in this area and define key factors that affect how it is perceived by selected groups.

4. Conclusions

The results of the survey point to dissatisfaction with the overall tax and customs system in the Slovak Republic and its current state, with the most critical point being the labor costs dealing with the tax and customs agenda, as well as the time costs for processing this agenda. The findings and conclusions of the survey further point to the need to simplify the system of taxes and duties in the form of an electronic agenda and emphasize the need for transparency of public finances. We perceive a general distrust of citizens towards the mechanism against tax evasion and the amount of money spent on tax and customs offices. Easier communication of the tax subject with a single point of contact has the prerequisites to eliminate uncertainty and to create an equivalent, free relationship between the citizen and the country, in which it is possible to build a positive image of the tax administration. This ultimately increases the percentage of voluntary tax performance (in the interest of the Ministry of Finance of the Slovak Republic, 96% is voluntary performance), based on a sense of belonging and responsibility to the state. This will also guarantee a feeling of fair distribution and thus a healthy, functioning economic system in Slovakia. From a managerial point of view, in line with the initial statement that the bureaucracy should be an executive apparatus taking over principles, relationships, and processes from the private sector, the following expectations are important:

-

better management of the state’s claims, with the possibility of offsetting them and a stronger bankruptcy position

-

use of the powers of the Financial Criminal and Judicial Office for the investigation of the tax and customs offenses

-

optimizing the number of employees by eliminating the duplication of effort and reducing management posts

-

reduction in the running costs of the merged organization

Through a concerted effort, it is possible to transform the tax and customs administration into an institution that will serve as a model for other parts of the public administration in the approach to the citizens so that the citizen-to-state financial relationship and related aspects are no longer perceived as negative but as a partnership that benefits both parties. Last but not least, we would like to state that it is necessary to make an enormous effort to combat tax fraud by measures that are listed in the material of the Ministry of Finance of the Slovak Republic as a draft concept of combating VAT fraud. This activity needs to be globalized within the EU. If the EU does not take serious and consistent action, such as binding materials, then we will see the possibility of eliminating tax fraud in the abolition of VAT and imposing an indirect sales tax, as is the case in some developed countries of the world.

The obtained research results, indicating dissatisfaction with the overall tax and customs system in Slovakia and the most critical tax collection processing issues, provide material for future observation.