Sustainable Development of the Economy—A Case Study of the Impacts of the Size of Enterprises and Factors Affecting Performance

Abstract

Approaches to the sustainable development of enterprises are a continuous concern of EU countries, contributing to the achievement of national well-being. Companies today face the combined effects of a multitude of factors that affect their results. This study was carried out to analyse the factors that influence the enterprises in the non-financial sector (industry, trade, and services). The analysis used the related data from eight Eastern European countries in the period of 2018–2020, and includes companies by country, size, number of employees, number of companies by number of employees, turnover, added value and average productivity per employee in generating turnover and added value, respectively. To carry out the study, four research questions and four possible hypotheses were used. For data analysis, generalized linear models were used, and four models were obtained and statistically validated. The obtained results led to the identification of the factors associated with the dependent variables that were the basis for the creation of the models.

1. Introduction

In recent years, global business has faced numerous crises (for example, the COVID-19 pandemic and the Russian–Ukrainian war), which have created growing imbalances and instability. The Sustainable Development Goals (SDGs) are increasingly considered a framework for businesses to adapt and meet these challenges with viable solutions [1,2]. The crucial role of companies in achieving sustainable goals is underlined by the UN, thus: “No matter how big or small, and regardless of their industry, all companies can contribute to the SDGs” [3]. The efforts of the European Union (EU) and the United Nations, aimed at ensuring the economic integration of member countries through common policies, have led to convergence, sustainable development, and the promotion of harmonious social progress, economic growth, and environmental conservation [4,5]. Establishing a sustainable Europe and achieving predetermined sustainability goals are priorities in the individual policies of European countries [6,7,8]. However, structural reforms have certain objectives and features that vary from one country to another due to political regime differences, which produce differences in the activity of economic agents [9]. From an economic point of view, each country faces various aspects of economic agents (regime of production, employment, industrial relations, etc.), depending on the European social model of which the country is a part, but also on the effects of structural reforms on economic performance and environment [10,11]. The pursuit of sustainable development has become necessary in the field of corporate life [12] because companies exert a great influence on generating sustainability and cultivating practices designed to support them in achieving performance [12,13]. From the studies of Claro and Esteves, economic, environmental, and social concerns manifest themselves in different proportions and degrees from one country to another [14], and responsible investment in ESG [15] and the quality of governance [16,17] can bring benefits to economies. To promote their contributions to national and global sustainability, the dimensions of sustainability must be translated through business models and strategies [13,17].

Thus, the concern of large companies to develop sustainable development practices [18,19] can be observed, while SMEs consider them difficult to implement [20,21]. A more favourable legislative framework, due to the size of SMEs (number of employees and operations undertaken), often favours a lack of involvement [20], and some studies emphasize the direct link between success and the introduction of sustainable development practices [22].

Considering the conceptual, empirical, and contextual developments, it can be considered that the present research adds to existing studies, showcasing, on one hand, the economies (under the imprint of transformations generated by the steps towards sustainable economies). On the other hand, the variables that influence the performance of the economies (being countries with emerging economies), which seem to currently not be researched, are approached here.

The motivating research questions in the presented empirical investigation are as follows:

RQ1: What are the variables that influence turnover at the country level?

RQ2: What are the variables that influence added value at the country level?

RQ3: What are the variables that influence the average productivity of each employee in generating turnover at the country level?

RQ4: What are the variables that influence the average productivity of each employee in the generation of added value at the country level?

This paper contributes to the specialized literature, advancing and facilitating the discussion of what is known in this field through a synthesis [23,24]. Through a bibliometric analysis [25,26], this study locates trends in business research of the factors that influence the performance of economies through the prism of enterprises structured by size and country. Content analysis is the key [23] to explore the context of the economy’s performance from several perspectives: turnover, value added, average productivity of each employee in generating turnover, and generating added value. The contribution to the specialized literature is achieved by developing a conceptual framework supported by a multitude of empirical evidence, which illustrates the interconnections between factors, dimensions, and outcomes in the economy of countries. Thus, the contribution of enterprises to the economy is captured in the context of national and international approaches to sustainability through the lens of the SDGs.

Finally, this research opens new research opportunities and recommends areas for further investigation, allowing researchers to position the desired contributions in the field [25,26] for the advancement of knowledge in this critical field in new and meaningful ways [23,24].

2. Literature Review and Hypotheses Development

2.1. Entrepreneurial Innovative Orientation

In the current context of economic and social complexities, and the scarcity of natural resources [27,28], it becomes imperative for businesses to develop innovative solutions that prioritize sustainability goals [29]. This approach allows companies to reorganize, reinvent themselves, and cooperate with various external partners to acquire and implement technological innovations, including digital ones [30,31], and thus obtain operational advantages. On the other hand, the dynamic development of information and communication technology (ICT) has caused changes in almost all areas of life [32,33]. Therefore, digital solutions related to the concept of “Industry 4.0” and the innovations associated with them are topics of interest for many researchers, these being considered vital for companies in the current context of sustainable development [34,35].

Currently, the main objectives of the EU in relation to SMEs foresee requirements for the introduction of digital solutions to improve the level of digitization, maintain the principles of sustainable development, and improve access to the free market and sources of finance [36]. The implementation and use of modern and digital technologies in SMEs is a great challenge due to the limited resources they face [37]. Several authors believe that SMEs’ access to this type of resources will allow the exchange of ideas and solutions and will provide the opportunity to obtain the necessary solutions [38] and to make their own solutions available [39,40,41]. There are also studies related to the effect of increasing production efficiency through the spillover of knowledge from technological innovation in the agricultural sector, which is contradictory, and thus a negative effect can be seen among American SMEs, and a positive effect among European ones [42,43].

The importance of innovation and its connection with entrepreneurial orientation, collaborative innovation, and innovation performance is repeatedly emphasized in the specialized literature [44,45]. Entrepreneurial orientation, orientation to new markets, orientation to techniques, and organizational routines are considered approaches for innovation in an organization [46], which will lead to increased innovation skills and improved performance for consumers and markets [47,48]. There are also studies that show the direct links between entrepreneurial orientation, SMEs performance, and organizational culture, emphasizing the role of the latter as a mediator for achieving success [49].

Although the study carried out in this paper did not measure the innovative entrepreneurial orientation (due to the lack of data), it can be considered that the studies mentioned above present visible advantages, thus offering a possible explanation for the differences that exist between companies and between the results obtained at the macroeconomic level. The effect of entrepreneurial orientation on performance can be deemed beneficial, even though it cannot be measured in monetary terms.

2.2. Performance Orientation

In the highly extremely competitive global economy, small- and medium-sized enterprises (SMEs) often occupy an increasing proportion of global enterprises and contribute greatly to economic development in a growing proportion among global enterprises, contributing to the development, growth, and recovery of the economies of many countries [50]. SMEs are considered to contribute to economic recovery and growth [38,39], and encourage competitiveness, innovation, and creativity [51,52,53], thus becoming essential blocks in a country’s industrial and economic development [54]. According to the resource-based approach, SMEs are more easily oriented to the market. Therefore, using the entrepreneurial spirit can determine sustainable competitive advantages [51,52] that lead to performance, development, and economic growth [53,54]. Being an engine of economic growth and development [4,5,6] that is flexible and capable of adjustment [55,56], the SMEs can generate innovation and advanced production capabilities [57] and thus can contribute to increasing the performance of economies of which they are a part. Through the introduction of innovation, direct and indirect financial results appear in the form of savings (of costs) and then the potential for development, market, and new opportunities [58,59], which are considered positive effects aimed at improving the company’s ability to create performance through the results obtained [60,61].

The current challenge for SMEs is how increased sustainability leads to increased profitability among businesses [62]. According to specialized studies and regulations in the field, sustainability aims for companies to make progress in all three perspectives of sustainability (economic, environmental, and social) [63]; this also includes corporate social responsibility (CSR) and environmental, social, and governance (ESG) reporting [64]. Thus, Rahi et al. [65] found positive and negative relationships among SMEs related to sustainability practices and financial measures, while Cerciello et al. [66] points out that emphasizing sustainability can initially have a negative effect on company profits. Complementing these studies, one should note that positive results are achievable only from a certain level of maturity in terms of sustainability aspects, where the three dimensions (environmental, economic, and social) converge towards a balance [67,68].

Day by day, it is apparent that challenges among companies are becoming more and more complex, and the need for performance is designed by connecting production to market requirements [69] and by adapting to sustainability, and permanent improvement becomes essential for survival [68,70]. However, this realization requires dedicated time and effort for both management and operational activities [69]. The organization of processes and operations, in addition to providing process support [70] to facilitate compliance with environmental standards and regulations [71], provides a solid basis for decision-making [37], and the quality of products and services cannot be obtained without sustainable development [62].

In agreement with the aforementioned studies, the present paper focuses only on the analysis of business performance through the measurable component in monetary terms, given by the revenues obtained by the economies of the countries studied, without denying the importance of factors that cannot be measured in monetary terms. Thus, the turnover (income) obtained at the level of the economy of the studied countries is analysed through the prism of evolution and influencing factors. The influencing factors studied are the country, the size of the company, the number of employees, productivity, added value, etc. The “turnover” indicator was chosen because of the belief that it best reflects the operational performance produced at a macroeconomic level (through the total volume of enterprises) under the influence of all factors that cannot be measured monetarily (sustainability, changes in the organization of management and operations, continuous adaptation for competitiveness, etc.). Based on this fact, the following hypothesis is proposed:

H1: 

The turnover depends on the country and the number of enterprises.

2.3. Added Value Orientation

The change in the interrelationship between labour and physical capital has become an essential factor affecting the value creation of enterprises [72]. With the digital economy, digital information and knowledge have become key factors of production, and the composition and interrelationship of factors of production involved in value creation have changed again. The involvement of external stakeholders (suppliers, customers, investors, authorities, and so on), and adaptation combined with transformation are considered essential for the implementation of sustainability efforts [35] and value creation at the company level [72]. Xu et al. suggests that each member of the supply chain, if they adopt environmentally conscious practices, will in turn influence other members to follow suit [73]. Thus, companies in their dual role of suppliers and partners in the supply chain will send signals towards the combined encouragement of sustainability efforts [74], and then the value of the firm will increase because of increased performance [75].

The effort for sustainable economies is also given by the introduction of technological innovations, which allowed the digitization of financial services and thus increased financial inclusion, financial stability, and security, increasing economic development through the newly created value [76,77]. Some studies show that the added value obtained from production negatively influences economic growth [78,79], and others contradict this by emphasizing the advanced refinement of products and services [79]. Added value is also discussed through the lens of gross capital formation that stimulates production by increasing the quantity and quality of assets; subsequently, high production and productivity stimulate economic growth [80], and innovation and new technologies add new gains [81].

According to the previously mentioned studies, it can be considered that added value accumulated in the economy is the result of efforts made by the companies in that economy to create the capital necessary for development in the existing conditions (adaptation to sustainable economies, stability, financial security, etc.). Added value is thus the object of research and analysis within this study through the lens of evolution and influencing factors, because it is what remains at the disposal of the company after covering the cost of the factors involved in production and contributes, as working capital, to the financial performance of the company and of the economy. The influential factors considered in the study are the country, company size, year, number of employees, average productivity of each employee in generating added value, etc. Based on this fact, the following hypothesis is proposed:

H2: 

The added value depends on the country and the number of employees.

2.4. Productivity Orientation

Reforms in Eastern Europe show variability between countries in terms of market structures and firm productivity [82]. Thus, increased interest in studying the determinants of productivity at the firm level [83,84] can be observed, as productivity growth is seen as a driver of prosperity in Eastern Europe [85]. Then, joining the European Union (EU) produced beneficial effects for Eastern European companies, helping them increase their productivity [86,87] and achieve higher growth rates [88]. Research shows that firms can increase productivity by gaining access to inputs greater than those available in their home country [82] but also at better costs and/or quality [89]. Although all regions differ, it is noted that one area may be strategically more advantageous (or problematic) than another [90]. Factors such as the impact of cultural dimensions [91], the importance of the host country context [92], and even the consequences of location choices for shaping international sourcing strategies [89] are always discussed as factors that can influence performance and productivity.

The increase in industrialization and urbanization, as well as the transfer of some industries to developing countries, have caused pressure on the environment in these countries [93]. Thus, economic growth at the country level must be approached both from sustainable [94] and financial perspectives (through various rates of growth in total productivity, factor productivity, and labour) to reflect the level of industrial transformation and modernization [95]. Firms adopting sustainability practices expect to improve their reputation, competitive advantage, productivity, revenue, and innovation abilities [96]. They also improve their economic performance [97] and financial performance [98] through sustainability disclosures. As a result, employee satisfaction and organizational performance can improve [99], and companies with excellent sustainability performance exhibit a distinct organizational culture [100,101].

The present study did not allow measurements of the performance among the economies in terms of productivity from a sustainable perspective (due to the lack of data) but only the average productivity of the turnover and that of the added value at the level of the economy of each country under study, expressed in monetary terms. Considering the perspectives arising from the specialized literature, it is considered that the analysis of the average productivity of turnover and added value is important, because it allows measurements and comparisons of the performance of the countries under study, and consequently a discussion of the performance of the economies through the prism of the factors of influence.

Based on these facts, the following hypotheses are proposed:

H3: 

The average productivity of each employee in generating turnover is dependent on the size of the company and the number of employees.

H4: 

The average productivity of each employee in generating added value is dependent on the number of enterprises and the number of employees.

The purpose of this study is to analyse the evolution of countries through the prism of the indicators studied and to identify the variables that influence these indicators and that produce effects on the performance of the economies. To achieve this goal, we sought answers to the research questions and formulated models for the working hypotheses that check the variables included in the model if they are statistically validated.

Thus, the study was carried out for eight countries selected from Eastern Europe (former communist countries); the economies are comparable due to the similar conditions given by the political governance before the 1990s. Data for these countries include enterprises in the non-financial sector (industry, trade, and services) presented by enterprise size classes. The study shows the influencing factors (variables) for the following selected indicators within the economies of each country studied: turnover, value added, the average productivity of the turnover generated by each employee, and the average productivity of the added value generated by each employee. The analysis of influencing factors was carried out using a statistical analysis of linear inferences (interdependencies) using regression analysis with the use of the general linear model (GLM).

3. Materials and Methods

In this study, an analysis was carried out based on the data provided by the Eurostat database [102] for the years 2018–2020 (see Appendix A—

Table A1. Initial data for NE (2018) [102].

NE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	690	4331	8955	14,127	315,666
CZ	1654	7057	12,905	19,858	1,001,856
EE	165	1063	2125	3466	70,680
LT	367	2197	4349	6997	197,201
HU	941	4515	9608	18,425	566,058
PL	3364	15,474	31,111	51,598	1,858,814
RO	1662	7955	16,447	27,196	448,714
SK	599	2567	4319	7019	479,132

,

Table A2. Initial data for NE (2019) [102].

NE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	696	4365	8896	14,651	320,059
CZ	1663	7057	13,187	19,822	1,017,047
EE	163	1049	2146	3622	75,283
LT	376	2238	4422	7064	205,806
HU	961	4575	9715	18,828	613,012
PL	3290	14,961	31,700	54,150	1,918,147
RO	1652	7774	16,878	27,529	462,870
SK	594	2505	4307	6864	497,812

,

Table A3. Initial data for NE (2020) [102].

NE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	650	4103	8387	13,647	314,440
CZ	1614	6812	12,963	19,541	1,027,516
EE	150	1041	1979	3616	78,694
LT	372	2169	4322	6916	214,520
HU	929	4269	9106	18,109	636,383
PL	3251	14,712	31,318	53,035	1,963,893
RO	1574	7322	16,040	26,756	489,452
SK	573	2379	4101	6395	505,049

,

Table A4. Initial data for Turn (2018) [102].

T	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	45,573.50	34,270.40	18,339.40	11,870.40	31,098.70
CZ	245,240.70	114,515.70	57,138.70	31,767.40	92,424.00
EE	12,700.20	15,744.60			21,165.20
LT	31,012.00	21,823.20	13,295.40	8677.70	17,117.80
HU	139,049.20	65,976.30	32,683.40	24,832.90	61,768.60
PL	493,525.30	201,090.70	111,694.60	76,422.60	242,096.90
RO	138,994.30	64,558.50	35,715.40	25,612.30	49,052.40
SK	91,325.30	38,772.40	22,730.70	12,105.10	47,403.70

,

Table A5. Initial data for Turn (2019) [102].

T	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	48,352.0	37,138.2	19,395.9	13,176.9	33,869.7
CZ	257,486.4	112,835.5	62,574.0	32,439.1	96,506.6
EE	13,046.7	16,251.4	9850.4	7243.5	22,939.7
LT	33,363.2	23,376.2	13,768.9	9405.2	18,218.8
HU	151,521.6	67,232.5	34,646.2	24,186.6	63,799.1
PL	543,143.5	194,584.4	112,125.3	80,185.8	250,684.8
RO	147,203.6	69,678.3	40,773.1	25,513.9	55,445.8
SK	92,924.8	37,970.1	22,732.6	12,171.0	49,475.0

,

Table A6. Initial data for Turn (2020) [102].

T	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	44,404.0	35,061.7	20,195.4	13,064.7	35,408.8
CZ	234,978.2	103,299.4	55,241.0	29,890.5	90,017.1
EE	12,105.2	15,726.4	9017		20,053.9
LT	32,733.2	22,812.4	13,572.3		18,676.8
HU	143,722.9	56,110.9	32,256.0	24,464.7	63,526.5
PL	515,062.2	188,436.4	108,359.3	74,179.7	261,392.0
RO	139,829.4	67,480.2	39,545.6	23,944.9	56,890.6
SK	86,304.5	35,587.5	19,801.6	11,268.1	47,425.6

,

Table A7. Initial data for AddVal (2018) [102].

AV	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	10,557.2	6768.6	3516.9	2249.7	6177.5
CZ	49,321.7	23,488.8	10,411.0	6572.3	22,588.9
EE	3008.6	3583.1			3707.0
LT	7043.8	5266.8	2786.0	1760.6	4133.4
HU	32,089.7	12,409.8	6768.6	5571.4	14,647.3
PL	120,767.3	45,410.7	23,611.0	16,555.4	40,936.6
RO	34,225.2	13,754.8	7661.8	5712.0	13,325.3
SK	17,739.3	7450.4	3651.1	2209.4	9091.2

,

Table A8. Initial data for AddVal (2019) [102].

AV	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	11,775.3	7575.5	3847.7	2519.5	6790.8
CZ	51,916.5	24,400.0	11,118.0	6856.7	23,604.9
EE	3108.6	3774.6	2049.8	1437.0	4140.1
LT	8066.4	5742.5	2974.1	1872.2	4264.7
HU	33,318.9	13,151.2	7220.6	5216.6	16,145.1
PL	133,054.1	46,338.1	25,345.5	18,415.2	47,329.8
RO	36,265.8	14,478.9	9431.6	6108.0	16,199.8
SK	18,181.5	7779.2	3902.7	2189.3	9554.5

,

Table A9. Initial data for AddVal (2020) [102].

AV	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	11,790.1	7804.6	4047.9	2636.4	6964.8
CZ	49,795.6	22,827.5	10,941.5	6484.9	22,691.3
EE	2979.1	3879.2			3912.1
LT	8419.9	6122.5	3161.2	2039.4	4615.4
HU	31,367.4	12,090.2	6826.6	5071.2	14,279.7
PL	135,799.1	47,858.9	25,938.5	18,272.8	46,806.7
RO	36,265.8	14,478.9		5888.3	16,966.1
SK	18,061.3	7417.9	3689.4	2061.5	9962.2

,

Table A10. Initial data for PE (2018) [102].

PE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	526,660	430,079	269,166	188,914	600,160
CZ	1,254,557	722,209	382,297	272,233	1,147,031
EE	90,980	99,955			140,239
LT	264,267	212,815	132,152	93,389	285,072
HU	913,490	451,222	283,458	244,966	938,495
PL	3,235,909	1,598,486	900,797	739,465	3,347,389
RO	1,443,732	808,941	499,594	365,615	943,328
SK	466,196	261,814	131,434	94,518	698,576

,

Table A11. Initial data for PE (2019) [102].

PE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	522,392	432,588	266,467	195,340	601,106
CZ	1,247,186	723,048	390,597	271,393	1,157,246
EE	92,799	100,688	61,672	46,979	141,254
LT	274,888	218,659	133,856	94,230	298,394
HU	922,395	454,520	287,816	251,100	1,015,191
PL	3,392,078	1,538,294	915,268	771,692	3,415,842
RO	1,440,734	792,261	511,963	369,666	973,942
SK	455,938	256,453	131,725	93,009	716,208

,

Table A12. Initial data for PE (2020) [102].

PE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	498,227	405,062	250,793	182,819	580,270
CZ	1,204,255	695,900	384,792	268,761	1,156,567
EE	86,671	100,082			146,199
LT	272,156	213,634	131,289	92,555	310,337
HU	902,986	426,707	270,326	240,698	1,019,125
PL	3,334,374	1,519,964	908,465	756,282	3,477,828
RO	1,398,351	751,627	483,696	358,444	997,887
SK	439,771	244,366	124,893	85,273	709,745

,

Table A13. Initial data for AvProdTurn (2018) [102].

PE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	0.09	0.08	0.07	0.06	0.05
CZ	0.20	0.16	0.15	0.12	0.08
EE	0.14	0.16			0.15
LT	0.12	0.10	0.10	0.09	0.06
HU	0.15	0.15	0.12	0.10	0.07
PL	0.15	0.13	0.12	0.10	0.07
RO	0.10	0.08	0.07	0.07	0.05
SK	0.20	0.15	0.17	0.13	0.07

,

Table A14. Initial data for AvProdTurn (2019) [102].

PE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	0.09	0.09	0.07	0.07	0.06
CZ	0.21	0.16	0.16	0.12	0.08
EE	0.14	0.16	0.16	0.15	0.16
LT	0.12	0.11	0.10	0.10	0.06
HU	0.16	0.15	0.12	0.10	0.06
PL	0.16	0.13	0.12	0.10	0.07
RO	0.10	0.09	0.08	0.07	0.06
SK	0.20	0.15	0.17	0.13	0.07

,

Table A15. Initial data for AvProdTurn (2020) [102].

PE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	0.09	0.09	0.08	0.07	0.06
CZ	0.20	0.15	0.14	0.11	0.08
EE	0.14	0.16			0.14
LT	0.12	0.11	0.10	0.10	0.06
HU	0.16	0.13	0.12	0.10	0.06
PL	0.15	0.12	0.12	0.10	0.08
RO	0.10	0.09	0.08	0.07	0.06
SK	0.20	0.15	0.16	0.13	0.07

,

Table A16. Initial data for AvProdAddVal (2018) [102].

PE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	0.02	0.02	0.01	0.01	0.01
CZ	0.04	0.03	0.03	0.02	0.02
EE	0.03	0.04	0.03	0.03	0.03
LT	0.03	0.02	0.02	0.02	0.01
HU	0.04	0.03	0.02	0.02	0.02
PL	0.04	0.03	0.03	0.02	0.01
RO	0.02	0.02	0.02	0.02	0.01
SK	0.04	0.03	0.03	0.02	0.01

,

Table A17. Initial data for AvProdAddVal (2019) [102].

PE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	0.02	0.02	0.01	0.01	0.01
CZ	0.04	0.03	0.03	0.03	0.02
EE	0.03	0.04			0.03
LT	0.03	0.03	0.02	0.02	0.01
HU	0.04	0.03	0.03	0.02	0.02
PL	0.04	0.03	0.03	0.02	0.01
RO	0.03	0.02	0.02	0.02	0.02
SK	0.04	0.03	0.03	0.02	0.01

and

Table A18. Initial data for AvProdAddVal (2020) [102].

PE	Enterprise Size
	>250	50–249	20–49	10–19	0–9
BG	0.02	0.02	0.02	0.01	0.01
CZ	0.04	0.03	0.03	0.02	0.02
EE	0.03	0.04	0.03	0.03	0.03
LT	0.03	0.03	0.02	0.02	0.01
HU	0.03	0.03	0.03	0.02	0.01
PL	0.04	0.03	0.03	0.02	0.01
RO	0.03	0.02	0.02	0.02	0.02
SK	0.04	0.03	0.03	0.02	0.01

) of some economic indicators operating in the non-financial sector (industry, trade, and services). The analysis considers the processing of the data obtained from each country through the prism of the results they obtained as an outcome of the efforts of companies to implement the principles of sustainable development regulated by international bodies [4,103]. Information was collected for 8 countries, Bulgaria, Czech Republic, Estonia, Lithuania, Hungary, Poland, Romania, and Slovakia, according to the Eurostat methodology using surveys [104], which were indexed according to

Table 1. The country list, adapted from the Eurostat database [102].

Label	Abbrev. Country	Name of Country
c1	BG	Bulgaria
c2	CZ	Czech Republic
c3	EE	Estonia
c4	LT	Lithuania
c5	HU	Hungary
c6	PL	Poland
c7	RO	Romania
c8	SK	Slovakia

c—country.

.

These countries, which are part of Eastern Europe, are countries where the data can be compared both because of the positioning in a close geographical area and because of the historical past facing the transition from a centralized economy to a private market economy. To carry out this study, several indicators were used that classify the enterprises in the non-financial sector corresponding to the size class of the workforce, according to

Table 2. The situation of the initial indicators, adapted from the Eurostat database [102].

No.	Abbrv.	Unit	Indicator Explanations
1	NE	number	includes the number of active enterprises on each type of size class, depending on the number of employed persons
2	Turn	millions of euros	the total of all sales (without VAT) of goods and services carried out by enterprises
3	AddVal	millions of euros	the difference between the value of what is produced and the intermediate consumption that goes into production, less subsidies to production and costs, taxes, and fees
4	PE	number	the total number of people working in the various industries: employed and unemployed (e.g., family workers and delivery staff), except for agency workers.

NE—number of enterprises; T—turnover; AV—added value; PE—persons employed in the enterprises.

.

To analyse the indicators, an indexation was made of each size class of the enterprises, according to

Table 3. The size enterprise list, adapted from the Eurostat database [102].

Label	Number of Employees	Explanations for the Size of the Enterprises
s1	0 ÷ 9	enterprises with 0–9 employees
s2	10 ÷ 19	enterprises with 10–19 employees
s3	20 ÷ 49	enterprises with 20–49 employees
s4	50 ÷ 249	enterprises with 50–249 employees
s5	>250	enterprises with >250 employees

s—size class of the enterprise.

.

The 4 indicators were subjected to mathematical calculations to make the most of the analysis. The indicator “the average productivity of each employee in the generation of turnover” was calculated with the help of the indicator “persons employed in the enterprises” (PE), by reporting the values related to “turnover” to “persons employed in the enterprises” (PE), according to Equation (1):

$${AvProdTurn}_{s,c,y}=\frac{{Turn}_{s,c,y}}{{PE}_{s,c,y}}$$

(1)

where AvProdTurn_s,c,y—the average productivity of each employee in generating turnover associated with each size class of enterprises in the same country and in the same year; s—index of the size of enterprises, s = 1 ÷ 5 (see Table 3); c—index of the country, c = 1 ÷ 8 (see Table 1); y—year, y = 2018 ÷ 2020; Turn_s,c,y—turnover associated with each size class of enterprises in the same country and in the same year; PE_s,c,y—the persons employed in the enterprises associated with each size class of enterprises and for the same year.

The indicator “the average productivity of each employee in the generation of added value” was calculated by reporting the values related to “value added” to “persons employed in the enterprises”, according to Equation (2):

$${AvProdAddVal}_{s,c,y}=\frac{{AddVal}_{s,c,y}}{{PE}_{s,c,y}}$$

(2)

where AvProdAddVal_s,c,y—the average productivity of each employee in the generation of added value associated with each size class of enterprises in the same country and in the same year, and AddVal_s,c,y—the added value associated with each size class of enterprises in the same country and in the same year.

Later, the “share of turnover” indicator was processed starting from the “the total of all sales” indicator and transformed into the share of each turnover in the total turnover for each enterprise category, obtaining potential at country level, according to Equation (3):

$${ShTurn}_{s,c,y}=\frac{{Turn}_{s,c,y}}{\sum _{s=1}^5{Turn}_{s,y}}$$

(3)

where ShTurn_s,c,y—the share of turnover from each country by size class and year in the total turnover in the same size class and in the same year, Turn_s,c,y—the turnover associated with each size class of enterprises in the same country and in the same year, and ∑Turn_s,y—the total turnover associated with each size class of enterprises in the same year.

The indicator “the share of added value at the cost of factors” was calculated by means of the indicator “added value”, resulting in the share of added value in the total added value at the country level, according to Equation (4):

$${ShAddVal}_{s,c,y}=\frac{{AddVal}_{s,c,y}}{\sum _{s=1}^5{AddVal}_{s,y}}$$

(4)

where ShAddVal_s,c,y—the share of the added value from each country by size class and year in the total turnover in the same size class and in the same year, and ∑AddVal_s,y—the total added value associated with each size class of enterprises in the same year.

The indicator “share of average productivity of each employee in the generation of turnover” was calculated using the indicator “average productivity of each employee in the generation of turnover”, resulting in the share of average productivity of each country in the total value of the studied countries, according to Equation (5):

$${ShAvProdTurn}_{s,c,y}=\frac{AvProd{Turn}_{s,c,y}}{\sum _{s=1}^5{AvProdTurn}_{s,y}}$$

(5)

The indicator “share of average productivity of each employee in the generation of added value” was calculated using the indicator “average productivity of each employee in the generation of added value”, resulting in the share of average productivity of each country in the total value of the studied countries, according to Equation (6):

$${ShAvProdAddVal}_{s,c,y}=\frac{{AvProdAddVal}_{s,c,y}}{\sum _{s=1}^5{AvProdAddVal}_{s,y}}$$

(6)

The variables selected in the preliminary analysis are presented in

Table 4. The situation of the indicators studied in the preliminary analysis.

No.	Variable	Unit	Indicator Explanations
1	ShTurn	%	allows the analysis of countries from the perspective of the turnover achieved at the country level in the total turnover of the countries studied
2	ShAddVal	%	allows the analysis of countries from the perspective of added value, showing what remains at the end and contributes to the development of the economy
3	ShAvProdTurn	%	efficiency indicator showing production expressed in monetary units related on average per employee
4	ShAvProdAddVal	%	efficiency indicator that shows the added value produced and remaining after deducting consumption, reported on average per employee

.

Through Equations (1)–(6), the indicators that will be used in this study were defined and are presented in Table 4.

In the first part of the study, a preliminary analysis was carried out, which would allow the comparison between the countries studied for the analysed years and the selected indicators. This analysis will allow comparative results to be obtained at the level of the countries studied, showing the progress registered by the country as a result of the performance of the companies participating in the economy. A high turnover in relation to a small number of employees can indicate increased work efficiency and good management of human and material resources. The reverse effect indicates possible problems in operational efficiency or the need for restructuring or personnel optimization. At the same time, a high level of added value may indicate efficiency in the use of human resources and may reflect qualitative and efficient work, and a low level may indicate the need for improvements in production processes or personnel management.

Later, a statistical analysis was performed with the help of Statistica software (v. 8.0) to answer the research questions, and the formulated hypotheses were verified. Thus, 4 models were created in which all the variables were included (see

Table 5. The status of variables in the general linear model analysis.

No.	Variable	Meaning	Domain
			Category	Type
1	year	Years studied from 2018 to 2020	Discrete	Ordinal
2	size	The size of the enterprises (see Table 3)	Discrete	Multinominal
3	country	The countries, SK is taken as a reference and compared with other countries (see Table 1)	Discrete	Multinominal
4	NE	The number of active enterprises in each type of size class	Discrete	Ordinal
5	PE	The persons employed in the enterprises	Discrete	Ordinal
6	Turn	Turnover	Continuous	Ratio
7	AddVal	Added value	Continuous	Ratio
8	AvProdTurn	The average productivity of each employee in the generation of turnover	Continuous	Ratio
9	AvProdAddVal	The average productivity of each employee in the generation of added value	Continuous	Ratio

), and the influencing factors affecting the indicators presented in Table 4 were verified. To perform an analysis using the general linear model (GLM), the variables have been separated into categories (continuous and discrete) and type (ordinal, multinomial, and ratio), according to Table 5.

The GLM analysis involved the following steps (for each model studied):

• Step 1—determining whether the association between response and term is statically significant;
• Step 2—determining how well the model fits the selected variables;
• Step 3—determining whether the model meets the assumptions of the analysis.

To determine how well the model fits the selected variables, an analysis was carried out up to the full factorial analysis. Thus, the number of variables was too high in relation to the number of cases to perform a complete factorial analysis (7! > 120 = 5!). For this, the analysis was reduced to a full factorial analysis with 2nd order effects (Equation (7) below).

$$\hat{Y}=\alpha +\sum _{1\le i\le 7}{\beta }_i{x}_i+\sum _{1\le i<j\le 7}{\gamma }_{i,j}{x}_i{x}_j$$

(7)

where Ŷ = Ŷ(x₁, …, x₇) is the equation of the GLM model involving x₁, …, x₇ as independent variables and α is the intercept, β_i is the coefficient of the variable x_i, and γ_i,j is the coefficient of the multiplicative effect between the values of the variables x_i and x_j (with i and j taking the values 1, 2, 3, 4, 5, 6, and 7). Notice that the number of variables is high (1 + 7 + ${\mathrm{C}}_7^2=29)$, so that such a model in full is unstable and its parameters do not (usually) have statistical significance. Subsequently, terms without statistical significance were eliminated in several steps (1, (x_i)_1≤i≤7, (x_ix_j)_1≤i<j≤7 are the terms in Equation (7)).

To determine how well the model fit the selected variables (see Table 5), several tests of univariate significance were performed. Consequently, those variables that did not have a statistically significant “p” value (p < 0.05) were gradually excluded. When the value of p is lower than 0.05, the null hypothesis can be rejected, the coefficient of the independent variable can be different from 0 and the chosen regression model is considered valid. Finally, the model that verifies the relationship of dependence between the dependent variable and the rest of the independent variables was chosen.

The global effect produced by “turnover” compared to the variables presented in Table 5 was obtained using Equation (8):

$$Turn=\widehat{Turn }+\epsilon$$

(8)

where $\widehat{Turn}=\widehat{Turn}\left(x_1,x_2,\dots x_n\right)$—the dependence of the turnover on the independent variables builds using Equation (7) as skeleton. The expression $\widehat{Turn}\left(x_1,x_2,\dots x_n\right)$ will be constructed as a sum of terms in which a term can be formed from an independent variable or from the product of two independent variables: (x = 1 ÷ n) and Ԑ—the residual error associated with the model.

Testing the dependent variable “turnover” (according to Equation (8)) with the independent variables through GLM analysis ultimately allowed us to obtain the GLM1 model, which shows the variables whose influence is statistically significant and that can influence the increase or decrease in turnover. Those that did not obtain a p < 0.05 for a coefficient were considered not to have a statistically significant influence on the model and were excluded.

Equation (9) was used to obtain the global effect produced by “added value” compared to the variables presented in Table 5:

$$AddVal=\widehat{AddVal }+\epsilon$$

(9)

where $\widehat{AddVal}=\widehat{AddVal}\left(x_1,x_2,\dots x_n\right)$ is the dependence of the added value on the independent variables expressed using Equation (7) as skeleton.

Testing the dependent variable “added value” (according to Equation (9)) with the independent variables through GLM analysis finally allowed us to obtain the GLM2 model, which shows the variables whose influence is statistically significant and that can influence the increase or decrease of the added value. As with GLM1, the variables that did not obtain a p < 0.05 for a coefficient were excluded from the model.

The global effect produced by “the average productivity of each employee in generating turnover” compared to the variables presented in Table 5 was verified by Equation (10):

$$AvProdTurn=\widehat{AvProdTurn+}\epsilon$$

(10)

where $\widehat{AvProdTurn}=\widehat{AvProdTurn}\left(x_1,x_2,\dots x_n\right)$ is the dependence of the average productivity of each employee in the generation of turnover expressed in Equation (7) as a function of the independent variables.

Testing the dependent variable “the average productivity of each employee in generating turnover” (according to Equation (10)) with the independent variables through GLM analysis finally allowed us to obtain the GLM3 model, which shows the variables whose influence is statistically significant and that can influence the increase or decrease of the indicator “the average productivity of each employee in generating turnover”. Gradually, the variables that did not obtain a p < 0.05 for a coefficient were excluded from the model until the model that was statistically validated was obtained.

The global effect produced by “the average productivity of each employee in generating added value” compared to the variables presented in Table 5 was verified by Equation (11):

$$AvProdAddVal=\widehat{AvProdAddVal+}\epsilon$$

(11)

where $\widehat{AddVal}=\widehat{AddVal}\left(x_1,x_2,\dots x_n\right)$ is the dependence of the average productivity of each employee in the generation of added value expressed in Equation (7) as a function of the independent variables.

Testing the dependent variable “the average productivity of each employee in the generation of added value” (according to Equation (11)) with the independent variables through GLM analysis finally obtained the GLM4 model, which shows the variables whose influence is statistically significant and that can influence the growth or decrease of the “average productivity of each employee in the generation of added value” indicator. Gradually, the variables that did not obtain a p < 0.05 for a coefficient were excluded from the model until the model that is statistically validated was achieved.

To better explain the chosen models, the size of the effect and its strength on the “turnover” indicator, the following descriptive statistics were calculated: df (degree of freedom), SS (total sum of squares), MS (model sum of squares), F-value (for the F-test), and p-value. SS represents the total variation of the data relative to their means, and a large value of SS reveals that the data are more variable relative to their means. MS is the variation that can be explained by the independent variables included in the regression model. Hence, if the value of MS in relation to SS is higher, the model explains the data variation better. According to the analysis of variance associated with the estimated regression (ANOVA), the F-value (for the F-test) is the test statistic used to assess the overall significance of the regression model. The F-test compared the variance explained by the model (SS) with the unspecified variance or residual variance (SR—residual sum of squares), reporting them and adjusting for the number of parameters estimated in the model. Higher values of the F-value statistic indicate that the model fits the data better. The p-value associated with the F-value must be less than the associated threshold (of 0.05), so that the null hypothesis—that all regression coefficients are zero—is rejected, and then the regression models are significant.

4. Results and Discussion

4.1. Preliminary Analysis

A comparative analysis of the countries in 2020 was carried out between the turnover and the average productivity of each employee in the generation of the turnover (see Appendix A—Figure A1a,b, and between the added value and the average productivity of each employee in the generation of added value (see Appendix A—Figure A1c,d).

Comparing the share of the turnover with the share of the average productivity of each employee in the generation of the turnover, at the level of 2020, the results are different at the level of the countries. Therefore, it can be said that the average productivity of each employee is one of the characteristics that makes the difference between the incomes within each economy. It can be observed that PL, CZ, HU, and RO are in the first place with the highest share of turnover in the figure of total business for approximately all size categories of enterprises (see Appendix A—Figure A1a). When the number of employees contributing to the turnover is also taken into account (see Appendix A—Figure A1b), one can notice that for s1 SK (17%), CZ (16.91%), and HU (13, 79%); for s2 EE (15.88%), CZ (15%), and SK (14.71%); for s3 SK (16.64%), EE (15.35%), and CZ (15.07%); for s4 EE (17%), SK (16.08%), and CZ (13.53%), and for s5 EE (23%), which is the country that distanced itself more than the countries in the same category (about 11%), followed by CZ (13.03%) and PL (12.58%).

The analysis of the added value indicator, compared to the share of the average productivity of each employee in the generation of added value, shows the performances recorded by the economy of each country and the efficiency of the use of production factors by each company (see Appendix A—Figure A1c,d). Thus, the share of the average productivity of each employee in the generation of added value expresses the result of the combination of production factors (material, human, and financial) in terms of efficiency at the level of each group of enterprises. From the analysis of the “share of added value” indicator, the first places are observed for PL, CZ, and RO with the highest share of added value in the total added value recorded by the eight countries (see Appendix A—Figure A1c). If the analysis continues, it can be noticed that the indicator “the share of the average productivity of each employee in generating of added value”, is more conclusive from the point of view of the result that remains in the company, which contributes to supporting the economy through taxes and fees and to supporting companies through the continuation of the activity and the possibility of new investments. Thus, in the following countries, in s1 the companies that are in first place are CZ (15.16%), SK (15.06%), and PL (14.93%); in s2, the companies are EE (16.93%) followed by CZ (14.33%) and PL (13.75%); in s3, we have EE in first place (16.3%), followed by SK (14.48%), and PL (14%); in s4, in first place is EE (17.28%), then SK (13.66%), PL (13.65%), and CZ (13.63%), and in s5, there are EE (20.31%), CZ (14.89%), and RO (12.9%). EE is the country producing the best added value, distancing itself by about 5% in the s5 category, 4% in the s4 category, and 2% in s2 and s3. This country has the best economic potential, obtaining the leading position in terms of cost management, in the sense of reducing them, and obtaining a surplus for all four types of enterprises located in categories s2–s5.

4.2. Results of the General Linear Model

Following the statistical analysis of the studied variables,

Table 6. Pearson’s correlation coefficients *.

Variables	year	NE	T	AV	PE	AvProdTurn	AvProdAddVal
year	1
NE	0.008319	1
Turn	0.0003372	0.2312	1
AddVal	0.02135	0.1593	0.9894	1
PE	−0.0006964	0.5840	0.8981	0.8655	1
AvProdTurn	0.004708	−0.3966	0.2755	0.2567	−0.03919	1
AvProdAddVal	0.08511	−0.4261	0.3713	0.3872	0.05236	0.9374	1

* Four significant digits are provided; in bold a high correlation is emphasized.

was obtained, which shows the degree of correlation (dependence) between the variables.

Analysing the results obtained in Table 6, the correlation between AvProdTurn and AdProdAddVal is high. According to this situation, AvProdTurn was removed from the explanatory variables and AvProdAddVal was kept at the level of the first model (GLM1).

Therefore, Table 6 highlights the linear associations and highlights a strong correlation between AvProdTurn and AdProdAddVal (approx. 0.94). This was considered for the other models, and one of these variables was eliminated to obtain significant models.

To see if there is a relationship between the country (the dependent variable) and the variable taken in the study (independent variables), the correlation coefficient (R), presented in

Table 7. The effect of variables: Test of the SS whole model vs. SS residual *.

No	Variable	Multiple		Adj. R2	SS	df	MS	Residual			F	p
		R	R2					SS	df	MS
1	Turn	0.9995	0.9990	0.9983	1.025 × 1012	46	2.2281 × 1010	1.067 × 109	68	1569 × 107	1420.5	0.00
2	AddVal	0.9999	0.9997	0.9996	5.841 × 1010	43	1.3583 × 109	1.489 × 107	71	2.098 × 105	6476.0	0.00
3	AvProdTurn	0.9998	0.9996	0.9989	1.779 × 101	72	0.002471	7.100 × 10−5	41	0.000002	1425.0	0.00
4	AvProdAddVal	0.9999	0.9997	0.9990	7.623 × 10−3	83	0.000092	2.000 × 10−6	30	0.000000	1384.0	0.00

R: Pearson’s correlation coefficient; SS: sum of squares; df: degrees of freedom; MS: mean of squares; F: F-value; p: p-value. * Four significant digits are provided.

, was calculated.

Analysing the results obtained in Table 7, there is a strong linear association relationship, since R² has values close to one for all variables. These results indicate that the values of the variables are strongly influenced by the country.

According to the GLM1 analysis, the resulting model that verifies the dependency relationship between the studied variables and between the dependent variable—turnover—and the rest of the independent variables, respectively, is presented in Equation (12):

$$\begin{gathered} Turn=a_1\times country+a_2\times size\times country+a_3\times size+a_4\times NE \\ +a_5\times AvProdAddVal+a_6\times size\times AddVal \\ +a_7\times year\times AvProdAddVal+\epsilon \end{gathered}$$

(12)

According to the GLM1 analysis (Equation (12)) and the data obtained from the method of least squares (see Appendix A—

Table A19. Parameter estimates for GLM1. Sigma-restricted parameterization for the parameter value Turn.

The Effect of Turnover	Level of Effect	Column	Value
			Param.	Std.Err	t	p
country	BG	1	−8197	1394	−5.87986	0.000000
country	CZ	2	11,298	2454	4.60302	0.000019
country	EE	3	−36,536	7406	−4.93317	0.000006
country	LT	4	−23,074	3664	−6.29775	0.000000
country	HU	5	−3741	1167	−3.20471	0.002060
country	PL	6	73,548	12,298	5.98037	0.000000
country	RO	7	7654	1175	6.51322	0.000000
size × country	1	8	−22,417	7774	−2.88340	0.005263
size × country	2	9	33,996	5744	5.91844	0.000000
size × country	3	10	−23,507	11,103	−2.11720	0.037904
size × country	4	11	−22,368	9083	−2.46261	0.016331
size × country	5	12	658	2310	0.28474	0.776706
size × country	6	13	47,795	29,215	1.63598	0.106465
size × country	7	14	−6804	2702	−2.51790	0.014166
size × country	8	15	3760	6422	0.58545	0.560181
size × country	9	16	−6852	4307	−1.59079	0.116296
size × country	10	17	1137	9039	0.12574	0.900310
size × country	11	18	4751	7838	0.60612	0.546453
size × country	12	19	3266	2107	1.55001	0.125781
size × country	13	20	−6880	25,118	−0.27392	0.784978
size × country	14	21	−105	2722	−0.03840	0.969481
size × country	15	22	28,392	10,346	2.74419	0.007751
size × country	16	23	−57,890	11,914	−4.85887	0.000007
size × country	17	24	60,530	17,565	3.44597	0.000980
size × country	18	25	40,979	12,889	3.17931	0.002223
size × country	19	26	750	2107	0.35606	0.722900
size × country	20	27	−98,142	38,849	−2.52623	0.013863
size × country	21	28	1747	7900	0.22109	0.825686
size × country	22	29	−13,598	6606	−2.05856	0.043371
size × country	23	30	26,634	7005	3.80192	0.000310
size × country	24	31	−34,716	9594	−3.61843	0.000565
size × country	25	32	−25,041	7863	−3.18450	0.002189
size × country	26	33	−3525	2536	−1.39037	0.168950
size × country	27	34	63,219	25,882	2.44258	0.017186
size × country	28	35	3188	3238	0.98451	0.328357
size	s5	36	16,699	12,147	1.37480	0.173707
size	s2	37	652	10,864	0.06001	0.952321
size	s1	38	−54,352	20,370	−2.66826	0.009524
size	s4	39	35,115	12,366	2.83958	0.005953
NE		40	0	0	3.59415	0.000611
AvProdAddVal		41	223,480,187	38,062,971	5.87133	0.000000
size × AddVal	1	42	2	0	7.09588	0.000000
size × AddVal	2	43	−2	2	−1.11661	0.268092
size × AddVal	3	44	−0	1	−0.01923	0.984713
size × AddVal	4	45	−0	1	−0.61699	0.539303
year × AvProdAddVal		46	−109,975	18,833	−5.83958	0.000000

), the value of the “Turn” indicator depends on the country, on the correlation between the size of the company and the country it belongs to, and on the longevity of companies by the average productivity of each employee in the generation of added value, the correlation between the size of the company and the added value, by the correlation between the year, and the average productivity of each employee in the generation of added value. Thus, the value of the coefficient “a”, which is a constant, indicates a positive direct connection (+), and the inverse when they have negative values (−) (see Appendix A—Table A19).

The coefficient “a₁” associated with “country” is significant in the model, as it has high positive values in the countries PL (73,548) and CZ (11,298), which show a direct dependence relationship (when turnover increases in SK (reference), it also grows in these countries). Large negative values are recorded in the countries EE (−36,536) and LT (−23,074), which show an indirect connection (when the turnover in SK increases, it decreases in these countries).

The coefficient “a₂” associated with “size × country”, where the reference for “size” is s₃ (20–49 employees) and the reference country is SK, is significant in the model, and has the following characteristics:

• For companies of size s₅ (>250 employees), large positive values are in PL (47,795) and it expresses that the turnover in this country increases a lot when it also increases in companies of the same size in SK (reference), and has negative values for EE (−23,507), BG (−22,417), and LT (−22,368), which reveals that the turnover decreases in these countries, while in SK it increases;
• For enterprises of size s₂ (10 ÷ 19 employees), higher positive values are in LT (4751), BG (3760), and HU (3266), with a direct link and negative values (inverse link) for PL (−6880) and CZ (−6852);
• For enterprises of size s₁ (0 ÷ 9 employees), higher positive values are in EE (60,530), LT (40,979), and BG (10,346), and negative ones are in PL (−98,142) and CZ (−57,890);
• For enterprises of size s₄ (50 ÷ 249 employees), higher positive values are in PL (63,219) and CZ (26,634), and negative ones are in EE (−34,716), LT (−25,041), and BG (−13,598).

The coefficient “a₃” associated with “size” is significant in the model, as it has large positive values in enterprises with 50 ÷ 249 employees (s₄: 35,115), in enterprises with 20 ÷ 49 employees (s₃: 16,699) and in enterprises with 10 ÷ 19 employees (s₂: 652), and large negative values in companies with 0 ÷ 9 employees (s₁: −54,352).

The coefficient “a₄” associated with “NE” is significant in the model, but the appropriate coefficient value of “0” reveals that it does not influence the model.

The coefficient “a₅” associated with “AvProdAddVal” is significant in the model, as it has a large positive value of ≈223 million, which shows that the turnover is highly dependent on this indicator.

The coefficient “a₆” associated with “size × AddVal” is significant in the model, as it has lower values; here, a direct connection with large companies (s₅) and an inverse relationship with the other companies (s₂, s₁ and s₄) can be seen.

The coefficient “a₇” associated with “year × AvProdAddVal” is also significant in the model, as it has a large negative value (−109,975). It reveals that the turnover indicator depends on the correlation “year × AvProdAddVal” and that the turnover will decrease when this indicator decreases.

According to the GLM1 model, the over-unit coefficients (>1) have a multiplicative role for the variable to which they are applied (see a₁, a₂, a₃, and a₇), and the subunit coefficients (<1) have a role of reducing the effect produced by the variable (see a₄, a₅, and a₆). Thus, it can be said that, according to the value of the coefficients (a₁, a₂, and a₃), the variables “country”, “size × country”, and “size” have the greatest influence on the increase in turnover. A decrease in turnover is found due to the negative values for certain coefficients (a₁, a₂, a₃, and a₇), and it shows that the variables “country”, “size × country”, “size”, and “year × AvProdAddVal” can also determine a multiplier effect of decrease for certain countries, size classes of enterprises, or because of the average productivity recorded in added value. Subunit values were obtained for coefficients a₄, a_5, and a₆ for the variables “NE”, “AvProdAddVal”, and “size × AddVal”, which indicate that the effect on turnover is lower for these variables.

The presence of this bivariate multinomial effect in the model is supported by specific national legislation that may favour certain size classes at the enterprise level, and there may also be differences from country to country.

Thus, for the restricted Sigma parameterization related to the dependent variable “turnover” and the independent variables presented in the model GLM1,

Table 8. Univariate tests of significance for GLM1: effect sizes and powers for turnover.

The Effect of Turnover	df	SS	MS	F	p
country	7	1.798429 × 109	256,918,379	16.3789	0.00000
size × country	28	7.640658 × 109	272,880,631	17.3966	0.00000
size	4	1.643784 × 108	41,094,598	2.6198	0.04238
NE	1	2.026298 × 108	202,629,771	12.9180	0.00061
AvProdAddVal	1	5.407324 × 108	540,732,366	34.4725	0.00000
size × AddVal	4	7.916015 × 108	197,900,375	12.6164	0.00000
year × AvProdAddVal	1	5.349000 × 108	534,899,971	34.1007	0.00000
Residual Error	68	1.066642 × 109	15,685,909
Total	114

was obtained.

According to Table 8, several observations (114) were made for each model, presenting the number of degrees of freedom in tables. Analysing the values obtained by “SS”, it can be said that within the GLM1 model, the largest total variation of the data in relation to their average is found in “size × AddVal”, “size × country”, and “AvProdAddVal”, and this reveals that the data are more variable in relation to their means. Large values recorded for “MS” compared to “SS” allow the conclusion that the model better explains the variation in the data. Higher values of the F statistic indicate that the model fits the data better.

Analysing Table 8, one can notice that the value of “p” is statistically significant for all the variables included in the model, which indicates that the model is valid, and thus an answer to the first research question H1 (What are the variables that influence turnover at the level of the country?) is obtained.

According to the GLM2 analysis, the resulting model, which verifies the dependency relationship between the studied variables, and between the dependent variable—added value—and the rest of the independent variables, respectively, is presented in Equation (13):

$$\begin{gathered} AddVal=a_1\times size\times country+a_2\times country\times year+a_3\times country\times PE \\ +a_4\times AvProdAddVal+\epsilon \end{gathered}$$

(13)

According to the GLM2 analysis (Equation (13)) and the data obtained from the method of least squares (see Appendix A—

Table A20. Parameter estimates for GLM2. Sigma-restricted parameterization for the parameter value AddVal.

The Effect of Added Value	Level of Effect	Column	Value
			Param.	Std.Err	t	p
size × country	1	1	2531.3	1141.881	2.21677	0.029843
size × country	2	2	71,85.8	3534.825	2.03284	0.045809
size × country	3	3	−65.2	256.377	−0.25444	0.799890
size × country	4	4	1493.2	1456.963	1.02490	0.308888
size × country	5	5	1801.8	2013.863	0.89469	0.373978
size × country	6	6	3126.7	3096.176	1.00986	0.315992
size × country	7	7	−15,035.7	3104.705	−4.84289	0.000007
size × country	8	8	−4133.4	1864.090	−2.21737	0.029800
size × country	9	9	−6497.9	3578.370	−1.81588	0.073610
size × country	10	10	541.9	876.927	0.61792	0.538603
size × country	11	11	−2242.7	2269.815	−0.98805	0.326485
size × country	12	12	−1265.8	1943.367	−0.65136	0.516918
size × country	13	13	−794.1	2909.952	−0.27289	0.785728
size × country	14	14	11,756.4	2297.890	5.11616	0.000003
size × country	15	15	3739.0	1733.801	2.15651	0.034432
size × country	16	16	4656.0	3105.569	1.49923	0.138247
size × country	17	17	−674.1	1054.950	−0.63901	0.524871
size × country	18	18	1827.8	2019.655	0.90500	0.368526
size × country	19	19	990.7	2403.518	0.41218	0.681453
size × country	20	20	−1843.2	3611.158	−0.51041	0.611349
size × country	21	21	−3998.3	906.592	−4.41023	0.000036
size × country	22	22	552.3	316.500	1.74514	0.085287
size × country	23	23	−404.1	394.386	−1.02457	0.309045
size × country	24	24	−128.7	340.937	−0.37743	0.706979
size × country	25	25	302.8	355.209	0.85235	0.396889
size × country	26	26	−415.8	771.342	−0.53906	0.591533
size × country	27	27	58.8	1053.067	0.05581	0.955649
size × country	28	28	−839.3	256.278	−3.27515	0.001635
country × year	1	29	3.9	1.745	2.24749	0.027717
country × year	2	30	5.0	2.784	1.79387	0.077092
country × year	3	31	−0.4	0.823	−0.53793	0.592310
country × year	4	32	2.1	2.090	1.01196	0.314992
country × year	5	33	1.1	1.656	0.65820	0.512541
country × year	6	34	0.6	2.308	0.24436	0.807658
country × year	7	35	−10.5	2.063	−5.07134	0.000003
country*PE	1	36	−0.0	0.009	−2.19335	0.031559
country × PE	2	37	−0.0	0.008	−1.50927	0.135669
country × PE	3	38	0.0	0.019	0.66162	0.510356
country × PE	4	39	−0.0	0.021	−0.94216	0.349304
country × PE	5	40	−0.0	0.006	−0.41096	0.682343
country × PE	6	41	0.0	0.002	0.38591	0.700715
country × PE	7	42	0.0	0.005	4.99810	0.000004
PE × AvProdAddVal		43	0.9	0.021	45.71024	0.000000

), one can observe that the value of the “AddVal” indicator depends on the correlation between the company size and the country it belongs to, on the correlation between the country and the year, on the correlation between the country and persons employed in the enterprises, and on the average productivity of each employee in the generation of added value.

The coefficient “a₁” associated with “size × country”, where the reference for “size” is s₃ (20–49 employees) and the reference country is SK, is significant in the model and has the following characteristics:

• For enterprises of size s₅ (>250 employees), large positive values are in CZ (7185.8), PL (3126.7), and BG (2531.3), and it shows that the added value increases a lot when it also increases in enterprises of the same size in SK (reference), and has negative values for RO (−15,035.7), which reveals that, in this country, the added value decreases, while in SK, it increases;
• For enterprises of size s₂ (10–19 employees), higher positive values are in RO (11,756.4) and EE (541.6), thus indicating a direct link, and there are negative values (inverse link) for CZ (−6497.9), BG (−4133.4), and LT (−2242.7);
• For enterprises of size s₁ (0–9 employees), higher positive values are in CZ (4656), BG (3739), and LT (1828), and negative values are in RO (−3998), PL (−1843), and EE (−674);
• For enterprises of size s₄ (50–249 employees), higher positive values are in BG (552), LT (302), and PL (59), and negative values are in RO (−839), LT (−416), and CZ (−404).

The “a₂” coefficient, associated with “country × year”, is significant in the model, and it has small positive values in CZ (5), BG (4), and LT (2), and negative values in RO (−10).

The “a₃” coefficient, associated with “country × PE”, is significant in the model, but has values close to zero and does not greatly influence the added value indicator.

The “a₄” coefficient, associated with “AvProdAddVal”, is significant in the model and it has a small value (0.9), which reveals that the added value is not greatly influenced by this indicator.

According to the GLM2 model, the super unit coefficient (>1) has a multiplicative role for the variable to which it is applied (see a₁), and the subunit coefficients (<1) have the role of reducing the effect produced by the variable (see a₂, a₃, and a₄). Therefore, it can be said that the value of the coefficient a₁ for the variable “size × country” has the greatest influence on the increase in the added value. A decrease in added value occurs when there are negative values for the coefficient a₁ and demonstrates that the variable “size × country” can also determine a multiplier effect of decrease for certain countries and the size class of the enterprise. Subunit values were also obtained for the coefficients a₂, a_3, and a₄ for the variables “country × year”, “country × PE” and “AvProdAddVal”, which show that, for these variables, the effect on the added value is diminished by the coefficient, thus having a lower influence.

The presence of this bivariate multinomial effect in the model is supported by the system of taxes and fees, which can favour certain size classes at the enterprise level. There can also be differences from one country to another as a result of the average productivity vis-à-vis the added value (and the cost of the factors, respectively), or the size of the enterprise as a result of the number of employees in the enterprise.

The restricted Sigma parametrization, which is related to the dependent variable “added value” and the independent variables verifying the model GLM2, was obtained in

Table 9. Univariate tests of significance of GLM2: effect sizes and powers for AddVal.

The Effect of Added Value	df	SS	MS	F	p
size × country	28	12,589,660	449,631	2.144	0.005231
country × year	7	5,998,065	856,866	4.085	0.000809
country × PE	7	5,887,828	841,118	4.010	0.000948
PE × AvProdAddVal	1	438,282,467	438,282,467	2089.426	0.00000
Residual Error	71	14,893,112	209,762
Total	114

.

According to Table 9, several observations (114) were made for each model, having presented the number of degrees of freedom in the table. Analysing the values obtained by “SS”, it can be said that within the GLM2 model, the largest total variation of the data in relation to their averages is found in “PE × AvProdAddVal” and “size × country”, and this reveals that the data are more variable in relation to their averages. The high values recorded for “MS” compared to “SS” allow the conclusion that the model better explains the data variation for PE × AvProdAddVal. Higher values of the F statistic indicate that the model fits the data better.

Upon analysing Table 9, the value of “p” is statistically significant for all the variables included in the model, which indicates that the model is valid, and thus an answer to the second research question H2 (What are the variables that influence the added value at the country level?) is obtained.

The GLM3 model, which verifies the dependency relationship between the studied variables, respectively, between the dependent variable—the average productivity of each employee in the generation of turnover—and the rest of the independent variables is presented in Equation (14):

$$\begin{array}{l}Val AvProdTurn=a_0+a_1\times size+a_2\times country+a_3\times size\times country+\\ a_4\times Turn+a_5\times PE+a_6\times country\times NE+a_7\times size\times Turn+a_8\times \\ country\times Turn+a_9\times size\times PE+a_{10}\times country\times PE+a_{11}\times NE\times PE+\\ a_{12}\times Turn\times PE+\epsilon \end{array}$$

(14)

According to the GLM3 analysis (Equation (14)) and the data obtained by the method of least squares (see Appendix A—

Table A21. Parameter estimates for GLM3. Sigma-restricted parameterization for AvProdTurn.

The Effect of the Average Productivity of Each Employee in Generating Turnover	Level of Effect	Column	Value
			Param.	Std.Err	t	p
intercept		1	0.092235	0.020041	4.60236	0.000040
size	s5	2	0.112263	0.032447	3.45983	0.001275
size	s2	3	0.011151	0.022507	0.49547	0.622913
size	s1	4	−0.155560	0.066418	−2.34213	0.024111
size	s4	5	0.019894	0.019848	1.00232	0.322069
country	BG	6	−0.017212	0.029635	−0.58079	0.564560
country	CZ	7	0.043032	0.027655	1.55602	0.127390
country	EE	8	0.035366	0.026742	1.32251	0.193327
country	LT	9	−0.090570	0.025483	−3.55408	0.000972
country	HU	10	0.010337	0.023308	0.44349	0.659741
country	PL	11	−0.019452	0.082534	−0.23568	0.814851
country	RO	12	−0.013855	0.020551	−0.67415	0.504000
size × country	1	13	−0.078508	0.036396	−2.15707	0.036914
size × country	2	14	0.060020	0.023615	2.54161	0.014911
size × country	3	15	−0.104027	0.030363	−3.42616	0.001404
size × country	4	16	−0.153863	0.036298	−4.23886	0.000124
size × country	5	17	−0.008653	0.020394	−0.42429	0.673570
size × country	6	18	0.351799	0.102935	3.41769	0.001439
size × country	7	19	−0.008580	0.022852	−0.37546	0.709257
size × country	8	20	−0.006532	0.023645	−0.27625	0.783745
size × country	9	21	−0.025472	0.019060	−1.33643	0.188777
size × country	10	22	−0.002846	0.021249	−0.13395	0.894101
size × country	11	23	0.023023	0.019409	1.18621	0.242369
size × country	12	24	−0.009767	0.012432	−0.78561	0.436614
size × country	13	25	0.044773	0.072112	0.62088	0.538116
size × country	14	26	0.001954	0.012179	0.16047	0.873295
size × country	15	27	0.117657	0.105961	1.11039	0.273303
size × country	16	28	−0.016675	0.057846	−0.28827	0.774593
size × country	17	29	0.134280	0.063218	2.12408	0.039742
size × country	18	30	0.204461	0.073676	2.77513	0.008274
size × country	19	31	0.024045	0.025783	0.93257	0.356502
size × country	20	32	−0.587630	0.271935	−2.16092	0.036595
size × country	21	33	0.019200	0.031384	0.61178	0.544058
size × country	22	34	−0.022107	0.028298	−0.78124	0.439148
size × country	23	35	0.001492	0.015243	0.09790	0.922490
size × country	24	36	−0.021544	0.019646	−1.09665	0.279196
size × country	25	37	−0.068091	0.022421	−3.03689	0.004144
size × country	26	38	0.001104	0.007165	0.15411	0.878280
size × country	27	39	0.131080	0.073125	1.79254	0.080427
size × country	28	40	−0.009445	0.010016	−0.94308	0.351163
Turn		41	0.000002	0.000000	17.28965	0.000000
PE		42	−0.000000	0.000000	−4.03290	0.000234
country × NE	1	43	−0.000000	0.000000	−0.32274	0.748535
country × NE	2	44	0.000000	0.000000	2.12613	0.039560
country × NE	3	45	−0.000000	0.000000	−0.38052	0.705522
country × NE	4	46	−0.000001	0.000000	−4.54205	0.000048
country × NE	5	47	0.000000	0.000000	2.13960	0.038388
country × NE	6	48	0.000001	0.000000	2.18501	0.034658
country × NE	7	49	0.000000	0.000000	1.20473	0.235214
size × Turn	1	50	−0.000001	0.000000	−8.02427	0.000000
size × Turn	2	51	0.000001	0.000000	4.79801	0.000021
size × Turn	3	52	−0.000001	0.000000	−4.59603	0.000041
size × Turn	4	53	−0.000000	0.000000	−1.34543	0.185880
country × Turn	1	54	0.000000	0.000000	1.63096	0.110557
country × Turn	2	55	−0.000001	0.000000	−5.01812	0.000011
country × Turn	3	56	0.000006	0.000001	10.07437	0.000000
country × Turn	4	57	−0.000004	0.000000	−8.71295	0.000000
country × Turn	5	58	−0.000000	0.000000	−1.78422	0.081789
country × Turn	6	59	−0.000002	0.000000	−7.61335	0.000000
country × Turn	7	60	−0.000001	0.000000	−5.26975	0.000005
size × PE	1	61	−0.000000	0.000000	−0.06650	0.947307
size × PE	2	62	−0.000000	0.000000	−2.68197	0.010500
size × PE	3	63	0.000000	0.000000	3.07140	0.003774
size × PE	4	64	−0.000000	0.000000	−0.06411	0.949193
country × PE	1	65	−0.000000	0.000000	−0.06236	0.950580
country × PE	2	66	−0.000000	0.000000	−0.72127	0.474836
country × PE	3	67	−0.000001	0.000000	−3.15418	0.003010
country × PE	4	68	0.000001	0.000000	6.26623	0.000000
country × PE	5	69	−0.000000	0.000000	−0.56733	0.573582
country × PE	6	70	0.000000	0.000000	1.40516	0.167507
country × PE	7	71	0.000000	0.000000	1.49778	0.141850
NE × PE		72	−0.000000	0.000000	−2.27556	0.028167
T × PE		73	0.000000	0.000000	3.36547	0.001669

), one can observe that the value of the “AvProdTurn” indicator depends on the intercept, company size, country, and the correlation between the company size and the country it belongs to; on turnover; on the number of employees’ on the correlation between the country and turnover; on the correlation between the country and number of enterprises; on the correlation between the country and number of employees; on the correlation between the company size and turnover; on the correlation between the company size and number of years—bets; on the correlation between the number of enterprises and the number of employees, and on the correlation between the turnover and the number of employees.

The “a₀” coefficient, associated with the intercept (the free term), has a value of 0.09.

The “a₁” coefficient, associated with “size”, is significant in the model, as it has low positive values in companies with >250 employees (s₅: 0.11), in companies with 50 ÷ 249 employees (s₄: 0.02), and in companies with 10 ÷ 19 employees (s₂: 0.01), and large negative values in companies with 0 ÷ 9 employees (s₁: −0.16).

The “a₂” coefficient, associated with “country”, is significant in the model and it has small positive values in the following countries: CZ (0.04), EE (0.04), and HU (0.01). This shows a direct dependence relationship (when it increases AvProdTurn in SK (reference), it also increases in these countries). Small negative values are recorded in the following countries: LT (−0.09), PL (−0.02), and BG (−0.02). This indicates an indirect connection (when AvProdTurn increases in SK, it decreases in these countries).

The “a₃” coefficient, associated with “size × country”, where the reference for “size” is s₃ (20–49 employees) and the reference country is SK, is significant in the model and has the following characteristics:

• For enterprises of size s₅ (>250 employees), small positive values are in PL (0.35) and CZ (0.06), which reveals that AvProdTurn increases when it also increases in enterprises of the same size in SK (reference), and has negative values for LT (−0.15) and BG (−0.08), which shows that AvProdTurn decreases in these countries, while in SK it increases;
• For enterprises of size s₂ (10–19 employees), lower positive values are in PL (0.4) and LT (0.2), a direct link, and negative values (an inverse link) are in CZ (−0.02), BG (0.01), and HU (0.01);
• For enterprises of size s₁ (0–9 employees), higher positive values are in LT (0.2), EE (0.13), and BG (0.12), and negative ones are in PL (−0.6) and CZ (−0.02);
• For enterprises of size s₄ (50–249 employees), higher positive values are in PL (0.13), and negative ones are in LT (−0.07), EE (−0.02), and BG (−0.02).

The “a₄” coefficient, associated with “Turn”, is significant in the model and it has small positive values (close to the value zero).

The “a₅” coefficient, associated with “PE”, is significant in the model, and it has small negative values (close to the value zero).

The “a₆” coefficient, associated with “country × NE”, is significant in the model, but values in all countries are close to the zero value.

The “a₇” coefficient, associated with “size × Turn”, is significant in the model, and it has lower values (closer to the zero value).

The “a₈” coefficient, associated with “country × Turn”, is significant in the model, and it has small values (close to the zero value).

The “a₉” coefficient, associated with “size × PE”, has values close to zero for all size classes of enterprises.

The “a₁₀” coefficient, associated with “country × PE”, is significant in the model, and it has values close to zero.

The “a₁₁” coefficient, associated with “NE × PE”, is also significant in the model, and it has a small value close to the zero value.

The “a₁₂” coefficient, associated with “Turn × PE”, is also significant in the model, and it has a small value close to the zero value.

According to the GLM3 model, all coefficients are subunits (<1), having the role of reducing the effect produced by the variable (see a₁–a₁₂). Thus, it can be said that the value of the coefficients allows the effect of the variable to increase when it is positive, and to reduce the effect when it is negative. The presence of this bivariate multinomial effect in the model is supported by the size class at the enterprise level, by the differences between countries at the level of sales as a result of the population and consumption possibilities, and by the number of employees in the enterprise.

The restricted Sigma parametrization, related to the dependent variable “the average productivity of each employee in the generation of turnover” and the independent variables that check the GLM3 model, is presented in

Table 10. Univariate tests of significance of GLM3: effect sizes and powers for AvProdTurn.

The Effect of the Average Productivity of Each Employee in Generating Turnover	df	SS	MS	F	p
Intercept	1	0.000037	0.000037	21.1817	0.00004
size	4	0.000021	0.000005	3.0232	0.02834
country	7	0.000053	0.000008	4.3637	0.001081
size × country	28	0.000493	0.000018	10.1455	0.000000
Turn	1	0.000518	0.000518	298.9319	0.000000
PE	1	0.000028	0.000028	16.2642	0.000234
country × Turn	7	0.000321	0.000046	26.4507	0.000000
country × NE	7	0.000044	0.000006	3.6379	0.003851
country × PE	7	0.000112	0.000016	9.2252	0.000001
size × Turn	4	0.000162	0.000040	23.3068	0.000000
size × PE	4	0.000025	0.000006	3.6133	0.01299
NE × PE	1	0.000009	0.000009	5.1782	0.028167
Turn × PE	1	0.000020	0.000020	11.3264	0.001669
Residual Error	41	0.000071	0.000002
Total	114

.

According to Table 10, several observations (114) were made for each model, presenting the number of degrees of freedom in tables. Analysing the values obtained by “SS”, one can conclude that within the GLM3 model, there is no large total data variation in relation to their averages. The values of “MS” compared to “SS” are not higher. Higher values of the F statistic indicate that the model fits the data better.

Upon analysing Table 10, it can be noticed that the value of “p” is statistically significant for all the variables included in the model, which reveals that the model is valid, and thus an answer to the third research question H3 (What are the variables that influence the average productivity of each employee in the generation of turnover at the country level?) is obtained.

The GLM4 model, which verifies the dependency relationship between the studied variables, namely, between the dependent variable “the average productivity of each employee in the generation of added value” and the rest of the independent variables, is presented in Equation (15):

$$\begin{array}{l}Val AvProdAddVal=a_0+a_1\times size+a_2\times country+a_3\times size \times \\ country+a_4\times AddVal+a_5\times PE+a_6\times country\times NE+a_7\times sizeTurn+\\ a_8\times country\times Turn+a_9\times NE\times Turn+a_{10}\times size\times AddVal+a_{11}\times \\ Country\times AddVal+a_{12}\times Turn\times AddVal+a_{13}\times size\times PE+a_{14 }\times \\ country\times PE+\epsilon \end{array}$$

(15)

According to the GLM4 analysis (Equation (15)) and the data obtained by the method of least squares (see Appendix A—

Table A22. Parameter estimates for GLM4. Sigma-restricted parameterization for AvProdAddVal.

The Effect of the Average Productivity of Each Employee in Generating Added Value	Level of Effect	Column	Value
			Param.	Std.Err	t	p
intercept		1	0.023774	0.004947	4.8058	0.000040
size	s5	2	0.024256	0.009757	2.4862	0.018708
size	s2	3	0.003005	0.005789	0.5191	0.607476
size	s1	4	−0.051907	0.014734	−3.5229	0.001390
size	s4	5	0.015650	0.005699	2.7459	0.010099
country	BG	6	−0.011632	0.005910	−1.9682	0.058347
country	CZ	7	−0.010765	0.007683	−1.4011	0.171453
country	EE	8	0.009455	0.006246	1.5139	0.140521
country	LT	9	0.004461	0.006634	0.6724	0.506458
country	HU	10	0.000563	0.004894	0.1151	0.909097
country	PL	11	−0.007448	0.022273	−0.3344	0.740406
country	RO	12	0.014087	0.006966	2.0221	0.052158
size × country	1	13	−0.037160	0.010122	−3.6713	0.000934
size × country	2	14	−0.010329	0.008075	−1.2792	0.210629
size × country	3	15	−0.023642	0.009010	−2.6240	0.013534
size × country	4	16	−0.016217	0.009654	−1.6799	0.103367
size × country	5	17	−0.006738	0.005471	−1.2316	0.227650
size × country	6	18	0.109679	0.031211	3.5141	0.001422
size × country	7	19	0.006781	0.005905	1.1484	0.259888
size × country	8	20	−0.000008	0.004793	−0.0017	0.998679
size × country	9	21	0.004351	0.004908	0.8865	0.382412
size × country	10	22	−0.005167	0.005423	−0.9528	0.348297
size × country	11	23	−0.005803	0.005158	−1.1249	0.269535
size × country	12	24	−0.001858	0.002776	−0.6693	0.508427
size × country	13	25	0.019777	0.019673	1.0053	0.322797
size × country	14	26	−0.006561	0.004020	−1.6321	0.113103
size × country	15	27	0.064812	0.023293	2.7824	0.009242
size × country	16	28	0.003508	0.012214	0.2872	0.775949
size × country	17	29	0.051597	0.013547	3.8087	0.000644
size × country	18	30	0.038046	0.015870	2.3974	0.022944
size × country	19	31	0.017099	0.009351	1.8285	0.077441
size × country	20	32	−0.229001	0.073311	−3.1237	0.003939
size × country	21	33	0.007546	0.009548	0.7904	0.435506
size × country	22	34	−0.021287	0.007033	−3.0270	0.005036
size × country	23	35	0.000061	0.003349	0.0182	0.985579
size × country	24	36	−0.013909	0.005215	−2.6672	0.012209
size × country	25	37	−0.008790	0.005323	−1.6513	0.109113
size × country	26	38	−0.004825	0.002603	−1.8533	0.073696
size × country	27	39	0.065608	0.021251	3.0873	0.004322
size × country	28	40	−0.004548	0.003045	−1.4937	0.145698
AddVal		41	0.000003	0.000000	18.6857	0.000000
PE		42	−0.000000	0.000000	−8.6793	0.000000
country × NE	1	43	−0.000000	0.000000	−2.2166	0.034379
country × NE	2	44	0.000000	0.000000	0.7951	0.432819
country × NE	3	45	−0.000000	0.000000	−1.8120	0.080002
country × NE	4	46	0.000000	0.000000	1.0693	0.293458
country × NE	5	47	0.000000	0.000000	0.8322	0.411893
country × NE	6	48	0.000000	0.000000	2.6803	0.011832
country × NE	7	49	0.000000	0.000000	1.9576	0.059640
size × Turn	1	50	−0.000000	0.000000	−4.3890	0.000130
size × Turn	2	51	0.000000	0.000000	1.1907	0.243125
size × Turn	3	52	0.000000	0.000000	2.4462	0.020515
size × Turn	4	53	−0.000000	0.000000	−2.7785	0.009331
country × Turn	1	54	−0.000000	0.000000	−1.7949	0.082745
country × Turn	2	55	0.000000	0.000000	0.3629	0.719193
country × Turn	3	56	−0.000001	0.000000	−3.7351	0.000786
country × Turn	4	57	0.000000	0.000000	4.5805	0.000076
country × Turn	5	58	0.000000	0.000000	3.4765	0.001572
country × Turn	6	59	−0.000000	0.000000	−0.3453	0.732266
country × Turn	7	60	0.000000	0.000000	4.6014	0.000072
NE × Turn		61	−0.000000	0.000000	−2.7193	0.010770
size × AddVal	1	62	−0.000001	0.000000	−6.4227	0.000000
size × AddVal	2	63	0.000001	0.000000	6.7650	0.000000
size × AddVal	3	64	−0.000001	0.000000	−6.1047	0.000001
size × AddVal	4	65	−0.000001	0.000000	−4.6322	0.000066
country × AddVal	1	66	0.000000	0.000000	0.7933	0.433839
country × AddVal	2	67	−0.000002	0.000000	−6.4810	0.000000
country × AddVal	3	68	0.000007	0.000001	8.0878	0.000000
country × AddVal	4	69	0.000001	0.000000	4.1209	0.000274
country × AddVal	5	70	−0.000002	0.000000	−6.5027	0.000000
country × AddVal	6	71	−0.000003	0.000000	−12.8922	0.000000
country × AddVal	7	72	−0.000003	0.000000	−13.6272	0.000000
Turn × AddVal		73	0.000000	0.000000	2.6195	0.013680
size × PE	1	74	0.000000	0.000000	2.3423	0.025991
size × PE	2	75	−0.000000	0.000000	−3.5473	0.001302
size × PE	3	76	0.000000	0.000000	4.9957	0.000024
size × PE	4	77	0.000000	0.000000	0.9765	0.336627
country × PE	1	78	0.000000	0.000000	4.4947	0.000097
country × PE	2	79	0.000000	0.000000	3.6299	0.001044
country × PE	3	80	−0.000000	0.000000	−3.0208	0.005115
country × PE	4	81	−0.000000	0.000000	−2.4057	0.022512
country × PE	5	82	0.000000	0.000000	1.5979	0.120553
country × PE	6	83	0.000000	0.000000	3.9243	0.000470
country × PE	7	84	0.000000	0.000000	1.8185	0.078989

), it can be observed that the value of the “AvProdAddVal” indicator depends on the size of the company, on the country, on the correlation between the size of the company and the country it belongs to, on the added value, on the number of employees, on the correlation between the country and the number of employees, on the correlation between the company size and turnover, on the correlation between the country and turnover, on the correlation between the number of companies and turnover, on the correlation between the company size and value added, on the correlation between the country and value added, on the correlation between turnover and value added, on the correlation between the company size and number of employees, and on the correlation between the country and number of employees.

The a₀ coefficient, associated with the intercept, has the value 0.02.

The “a₁” coefficient, associated with “size”, where the reference for size is s₃ (20–49 employees), is significant in the model and has the following characteristics:

• For companies of size s₅ (>250 employees), it has a small value (0.02) and reveals that the turnover is not influenced too much by this variable;
• For enterprises of size s₂ (10–19 employees), it has a small value (0.003), which does not influence the turnover too much;
• For companies of size s₁ (0–9 employees), it has a small negative value (−0.05);
• For enterprises of size s₄ (50–249 employees), a small number appears again (0.02).

The “a₂” coefficient, associated with “country”, is significant in the model, as it has positive values in the countries RO (0.01) and EE (0.01), which shows a link of direct dependence, and negative values in the countries BG (−0.01), CZ (−0.01), and PL (−0.01), revealing an indirect connection.

The “a₃” coefficient, associated with “size × country”, where the reference for size is s₃ (20–49 employees) and the reference country is SK, is significant in the model and has the following characteristics:

• For companies of size s₅ (>250 employees), positive values are in HU (0.01) and PL (0.01), which showcases the increase in turnover when it also increases in companies of the same size in SK (reference), and has negative values for the other countries (higher in BG (−0.04), EE (−0.02) and LT (−0.02)), which reveals that, in these countries, turnover decreases, while in SK, it increases;
• For enterprises of size s₂ (10–19 employees), higher positive values are in PL (0.02), a direct link, and there are small negative values (inverse link) in EE, LT and RO;
• for enterprises of size s₁ (0–9 employees), small positive values are in BG (0.06), EE (0.05), and LT (0.04), and negative ones in PL (−0.23);
• for enterprises of size s₄ (50–249 employees), small positive values are in PL (0.06) and negative ones are in BG (−0.02) and EE (−0.02).

The “a₄” coefficient, associated with “AddVal”, is significant in the model, and it has a small positive value.

The “a₅” coefficient, associated with “PE”, is significant in the model, and it has a small value close to zero.

The “a₆” coefficient, associated with “country × NE”, is significant in the model, and it has small values close to the zero value.

The “a₇” coefficient, associated with “size × Turn”, is significant in the model, and it has lower values, close to the zero value.

The “a₈” coefficient, associated with “country × Turn”, is also significant in the model, and it has lower values, close to the zero value.

The “a₉” coefficient, associated with “NE × Turn”, is also significant in the model, and it has lower values, close to the zero value.

The “a₁₀” coefficient, associated with “Size × AddVal”, is also significant in the model, and it has smaller values, close to the zero value.

The “a₁₁” coefficient, associated with “country × AddVal”, is also significant in the model, and it has lower values, close to the zero value.

The “a₁₂” coefficient, associated with “Turn × AddVal”, is also significant in the model, and it has smaller values, close to the zero value.

The “a₁₃” coefficient, associated with “size × PE”, is also significant in the model, and it has lower values, close to the zero value.

The “a₁₄” coefficient, associated with “country × PE”, is also significant in the model, and it has lower values, close to the zero value.

According to the GLM4 model, all coefficients are subunits (<1) having the role of reducing the effect produced by the variable (see a₁–a₁₄). Hence, it can be said that the value of the coefficients allows the effect of the variable to increase when it is positive, and to reduce the effect generated by it when it is negative. The presence of the bivariate multinomial effect in the model is supported by the size class at the enterprise level, by the country, by the differences between countries at the level of sales (because of population and consumption possibilities), and by the added value vis-à-vis the consumption of production factors and the number of company employees.

The restricted Sigma parametrization, related to the dependent variable “the average productivity of each employee in the generation of added value” and the independent variables that check the GLM4 model, is presented in

Table 11. Univariate tests of significance of GLM4: effect sizes and powers for AvProdAddVal.

The Effect of the Average Productivity of Each Employee in Generating Added Value	df	SS	MS	F	p
Intercept	1	0.000002	0.000002	23.0961	0.000040
size	4	0.000001	0.000000	4.1281	0.008797
country	7	0.000001	0.000000	2.5302	0.03590
size × country	28	0.000011	0.000000	5.7047	0.000005
AddVal	1	0.000023	0.000023	349.1570	0.000000
PE	1	0.000005	0.000005	75.3296	0.000000
country × NE	7	0.000001	0.000000	2.5445	0.03505
size × Turn	4	0.000001	0.000000	5.2885	0.002405
country × Turn	7	0.000004	0.000001	7.7656	0.000023
NE × Turn	1	0.000000	0.000000	7.3948	0.0108
size × AddVal	4	0.000007	0.000002	27.9915	0.000000
country × AddVal	7	0.000021	0.000003	45.6957	0.000000
Turn × AddVal	1	0.000000	0.000000	6.8617	0.013680
size × PE	4	0.000002	0.000001	8.8277	0.000078
country × PE	7	0.000003	0.000000	6.0456	0.000187
Residual Error	30	0.000002	0.000000
Total	114

.

According to Table 11, several observations (114) were made for each variable, with the number of degrees of freedom presented in the table. Analysing the values obtained by “SS”, it can be observed that within the GLM4 model, a large total data variation is not present. Additionally, the values recorded for “MS” compared to “SS” are not higher. Higher values of the F statistic indicate that the model fits the data better. Analysing Table 11, one can notice that the value of “p” is statistically significant for all the variables included in the model, which reveals that the model is valid, and thus an answer to the third research question H4 (What are the variables that influence the average productivity of each employee in the generation of added value at the country level?) is obtained.

The study by Nasrallah et al. [18], which analyses the impact of corporate governance in SMEs on performance, took a sample of 150 unlisted companies and identified the effect on the ROA and ROI performance indicators. Applying 2SLS regression to control endogeneity and a quantile regression, they studied the corporate governance score and how it affects each performance component. The present study confirms that company size affects performance, as an influential factor among companies, even if the analysis of Nasrallah et al. was realized from another perspective.

Ullah’s study [54] investigates SMEs from 28 countries in Eastern Europe and Central Asia through the lens of financial constraints (which affect sales and employment), and the effect of corruption in the business environment. Hence, with the help of some regression models, it is observed that countries with higher levels of economic, financial, and institutional development have fewer financial constraints, thus increasing sales and reducing corruption. Compared to the study presented here, Ullah’s study uses a different set of variables (firm constraints, firm size, firm age, exporter, foreign, government, ownership, privatized, manufacturing, and services), and the results are grouped by the effects of financial constraints. The present study uses several independent variables and offers, from another perspective, the factors that affect the growth of turnover, value added and average labour productivity for the generation of turnover and value added.

The study carried out by Gherghina et al. [105] uses several log-log linear regressions estimates for Romania, and shows the positive impacts of investments and innovation on the growth of the territorial economy, measured by turnover. It also reveals the positive impact of the company size on the turnover. This study is conducted only for Romania, while the present paper analyses eight countries. The variables used are the turnover compared to the investments made and the expenses for innovation (at the company size level). The present study has four dependent variables (turnover, added value, average productivity generated by turnover, and added value), which are compared with several 38 independent variables.

Using the linear regression model, the study by Nieto et al. [106] shows how SMEs can increase productivity. Therefore, in a sample of SMEs from Eastern Europe, he found a positive relationship between the firm’s productivity and inputs from regions with formal and informal institutional links. This study only addresses the topic of productivity according to other variables, while, in this paper, turnover and added value are also taken into consideration.

5. Conclusions

The evolution trend of enterprises in the economy of the countries studied is approximately linear, with no significant fluctuations in the three years studied. Upon analysing the turnover per employee in 2020, one can notice that the highest level is 23%, obtained by EE in small companies (0–9 employees), followed by a percentage of 17%, obtained by EE (at 10–19 employees), SK (20–49 employees), CZ, and SK (for enterprises with several employees (>250)). The turnover shows the ability of economy to assimilate (through sales) the goods and services provided by the companies, and thus restore the circuit of the resumption of the business activity.

Upon examining the indicator added value per employee in 2020, it can be noticed that the highest values are for small enterprises in countries such as EE, CZ, and RO, countries that obtain superior performances in terms of how they manage factor costs when there is a small number of employees. Being an indicator of financial performance, the added value is the one that indicates how big the gains are in the economy when the costs related to the factors have been reduced. At the level of all categories of enterprises in terms of the contribution to the achievement of added value, in 2020, EE (16.14%) was the country that ranked first, followed by CZ (14.43%), SK (13.72%), PL (13.64%), HU (12.16%), LT (11.89%), RO (9.6%), and BG (8.43%). The surplus value that the companies manage to generate in an economy, because of the effort put into the production and trade process, signifies the efficiency of the use of factors, on the one hand, and the progress recorded, on the other.

Although there are differences from one economy to another that are given by a multitude of variables, which can be measurable or immeasurable in monetary terms, it can be considered that the study carried out here allowed the identification of variables which can be statistically validated and that influence the performance at the level of the economies of the countries taken under study. Consequently, through the results of the indicators (turnover, added value, the average value of the turnover generated by an employee, and the average value of the added value generated by an employee), it was possible to obtain models that include the influencing factors on the performance in the economy.

The statistical analysis carried out using the general linear model showed that all the models obtained (for turnover (GLM1), added value (GLM2), the average productivity of each employee in the generation of turnover (GLM3), and the average productivity of each employee in value added generation (GLM4)) are valid. Through the statistical validation of these models, the research questions that were the basis of the study were answered, finding the influencing variables of the indicators under study. For each model, the variables that make up the model and allowed the verification of the hypotheses were obtained. Thus, hypothesis H1 (“Turnover depends on the country and the number of enterprises”) was verified by the GML1 model, including these influencing factors together with other factors, as well as their combined effect (see Equation (12)). Hypothesis H2 (“The added value depends on the country and the number of employees”) was verified through the GML2 model, and the influencing factors included the effect produced by the country and the number of employees (see Equation (13)), and observing, at the same time, the combined effects of these factors in correlation with other factors. Hypothesis H3 (“The average productivity of each employee in generating turnover depends on the size of the company and the number of employees”) was verified through the GML3 model, including these variables as influencing factors, along with other factors (see Equation (14)). Hypothesis H4 (“The average productivity of each employee in the generation of added value depends on the number of enterprises and the number of employees”) was verified by the GML4 model, including these influencing factors, along with other factors (see Equation (15)).

As a result of the above conclusions, it can be concluded that sustainable development within economies must also be supported through the prism of the factors that determine companies to achieve results. Thus, if a high added value at the country level is desired, the fact that it is obtained differently from one country to another must be considered, due to the internal conditions created by the policies implemented but also the number of employees and their contributions. The average productivity of each employee in generating turnover and added value exhibits, through the prism of the resulting model, the influencing factors that must be acted upon for sustainable development at the level of companies and the country.

6. Discussion

Referring to the preliminary analysis, this study shows that although the analysis was carried out on countries with emerging economies, there are countries where the efficiency indicators register increases (see Appendix A—Figure A1a: PL and Figure A1c: EE), and countries where they register stagnation or declines. This is also emphasized by the study by Lu et al. [107], who drew a parallel between groups of developed and emerging countries and showed that, at the EU level, the dynamic efficiency has increased in recent years by 4.23%. Referring to the productivity indicators, the study by Lu et al. [107] shows that productivity at the EU level has improved on the background of increased production and better employment of the labour force in developed countries compared to emerging ones. The present study adds to the existing ones, showing that the average productivity per employee increased only in a few countries, while in the majority, it decreased or stagnated (see Appendix A—Figure A1b: EE and Figure A1d: EE and CZ).

Referring to the influencing factors that affect companies, there are studies [18,54,105] that, like the present paper, include, among the significant variables, company size and country as factors affecting performance.

According to the studies of Nieto et al., the quality of the countries as EU members determines their formal institutional links that facilitate access to resources and can increase efficiency indicators, but the exposure and political impact (due to the communist regime) negatively and differently affect the countries and productivity gains. Agreeing with the above conclusions, the present study reflects a low efficiency and productivity in the years studied and the explanation through the influencing factors is considered opportune.

7. Practical Implications

Scientific curiosity led to the comparison of the differences between Eastern European countries, depending on the size of the companies and on the performance in the economy of each country. According to the review of previous studies, this is the first empirical study to test this issue in this context.

The present study also joins others that support the promotion of the entrepreneurial mentality [46], investments and innovation [100], access to resources, and the correct management of resources [101], and it can be argued that these factors lead to long-term development. Besides these factors, the differences between countries on a global level are also given by a combination of internal factors (economic and political) that determine differences in average productivity per employee both at the level of turnover and at the level of added values.

The study conducted serves various decision-makers in each country, who are taking steps towards the sustainable development of enterprises. The benefits can also be seen through the sustainability lens of companies that are continuously transforming towards implementing sustainability policies for long-term survival. Thus, sustainability must also produce positive effects on turnover, cost reduction to increase added value, increased productivity to generate turnover and added value, etc. Therefore, based on the outcomes obtained in the empirical research, the results can be used by decision-makers and practitioners in companies as follows: (1) governments can support and encourage companies towards sustainable development through regulations and subsidies that will have the effect of increasing performance and long-term savings, and (2) as owners or practitioners in companies, performance is expected to increase if the influencing factors are taken into account, contributing to the final long-term result. The present work is also useful for academic researchers to understand the differences between countries in terms of average productivity and the factors that contribute to the performance of economies and can be developed through new research approaches.

8. Limitations of the Study and Future Research

There are several limitations of the shown work, some of which may be further investigated in the future. Firstly, the presented model only investigates the situation of some countries in Eastern Europe, and the research can be extended to other countries as well. A possible extension would be to investigate how other countries have the same variables to which they refer or not. Secondly, this paper analyses the data without being able to associate the differences in values between countries with the socio-economic or political conditions that allowed these changes, which is impossible.