The Consequences of Economy Servitization for Ensuring Energy Sustainability—The Case of Developed and Developing Countries

Abstract

This study presents an analysis of the relationship between the servitization process and energy sustainability in the years 2015–2020. The research refers to 164 selected countries, also divided into two regimes: developed and developing. The transformation of the manufacturing process, and as a result, the economy’s structure, towards servitization, is observed in most countries worldwide. The positive influence of the servitization of production by individual manufacturers on sustainability is widely known. In this research, this relationship is considered on a macroeconomic scale, which is one of the novelties of the study. Particularly, sustainability in the energy sector, indicated as an achievement of the 7th goal of Sustainable Development, is discussed. Energy sustainability is evaluated using a synthetic measure by Perkal. This part of the research shows the problem of the low level of energy sustainability in developing countries (particularly in Africa) compared with developed ones. Moreover, spatio-temporal sensitivity models are estimated and verified. The sensitivity parameter in these models shows the impact of the progress in the servitization process on energy sustainability. The models have been enriched with the effects of spatial dependence between countries, taking into account two types of proximity matrices based on (1) the common border criterion and (2) the similarity of the development levels measured by the Human Development Index. Additionally, the differences in sensitivity between developed and developing countries are considered. The results of the study show that in both cases, the economic servitization positively influences energy sustainability, but the strength of the relationship is stronger in the group of developed countries. This can be, for example, the result of the individual characteristics of the given countries, where African countries mainly benefit from agricultural development. Only after reaching a certain level of economic growth will they be able to obtain sustainability faster through economic servitization.

1. Introduction

The progressing economic growth of countries requires an increase in the output level. The increase in total country production can be the result of growth in the agricultural, industrial, or service sectors of the economy. Nowadays, particular attention is paid to the development of the service sector. This trend is especially observed in developed countries [1] and is called servitization. The level of development determines the intensity of the servitization—the focus on the service sector increases with the increase in economic growth [2]. In the literature, servitization is defined, among others, as “the increasing offering of fuller market packages or ‘bundles’ of customer-focused combination of goods, services, support, self-service, and knowledge in order to add value to core corporate offerings” [3] or “the emergence of product-based service which blur the traditional distinction between manufacturing and traditional service sector enterprises” [4]. Moreover, companies start to use the approach named the product-service system (PSS), which provides an increase in consumer satisfaction while reducing the consumption of products [5,6].

The PSS has a positive impact on the environmental situation due to a reduction in the amount of materials used to satisfy human needs [7,8], which is opposite to the expansion of the industrial sector [9,10]. Moreover, the role of the servitization process in Sustainable Development (SD), especially in the environmental context, is strongly highlighted. The natural environment is threatened by the vast use of raw materials in the production process to meet unlimited consumer needs. Servitization allows suppliers to improve the durability of goods by changing the design of the products or business models, which leads to a reduction in environmental impact [11]. With the improvement in the environmental situation, the changes in energy consumption are also related. SD also considers energy sustainability.

The servitization process is closely linked with energy efficiency improvement. The focus on services in the production system leads to more efficient resource use and technological development and, as a result, enhances energy efficiency [12]. Moreover, companies can implement advanced service offerings due to the technological evolution related to the PSS, which also allows for optimizing the energy efficiency throughout the production process. The consequence of these changes is a weakening of the negative environmental impacts. This understanding of servitization provides insight into its relationship with sustainability in general [8,13,14], especially energy sustainability [15]. The importance of the relationship between economic development and the energy situation of the country is highlighted by Nassar et al. [16]. They also focus on the environmental determinants of renewable energy production, which also favor the improvement of social conditions. Hence, the linkage between economic, social, environmental, and energy sustainability is highly desirable. This study focuses only on two of these—economic development and energy sustainability. The remaining factors will be added in further research.

Energy sustainability is one of the crucial challenges for governments nowadays. In 2015, the United Nations (UN) General Assembly ratified the 2030 Agenda for Sustainable Development, which includes 17 goals. One of them concerns ensuring access to affordable, reliable, sustainable, and modern energy for all (SDG7). The objectives of monitoring SDG7 are as follows: (1) ensure access to modern energy services, (2) increase the importance of renewable energy consumption, (3) double the boost in energy efficiency, and (4) expand the infrastructure and raise the technology for supplying sustainable energy services in developing countries [17]. SDG7 particularly focuses on energy use and production in the environmental context, especially recognizing climate change [18,19].

This study concerns the relationship between economic servitization and energy sustainability. It has been carried out regarding the importance of the levels of development of the countries considered, and additionally, a spatial perspective was introduced. On the one hand, developed countries were selected for the study; on the other hand, this relationship was also examined for selected developing countries. In this context, the aim of the analysis is to assess the diversity of the relationship depending on the level of economic development. An additional objective of the study is to investigate the importance of geographical distance vs. economic proximity in the spatial analysis of the given relationship. The economic proximity is determined by the similarity in the Human Development Index (HDI) levels. In our study, we assumed that the servitization process is assessed by the changes in service value-added, and energy sustainability is determined by a synthetic measure based on the diagnostic variables characterizing the seventh goal of the SDGs.

2. Literature Review and Novelty of the Research

In the literature, servitization is perceived in two ways: (1) as a tool to improve customer loyalty and raise profits [20], or (2) as a process decoupling customer satisfaction from resource consumption [21]. Moreover, different terms used to characterize servitization include the transition from product to service, the product-service system, service infusion, and hybrid offerings [22]. All of these terms underline the importance of a service-centric approach to product distribution. Most researchers consider servitization from the producers’ point of view. They use measures such as the scale of service activities [23] or the ratio of service sales [24] to assess it.

Buera and Kaboski [25] pointed out the increasing importance of service consumption in economic growth. They conducted one of the few studies on the relationship between servitization and economic growth at the macro level. In turn, Capello et al. [26] discussed the size of a particular type of service, namely digital economy services, across the European regions. They explained the development of the digital services as a result of advancing technological expansion. Meanwhile, Dogan and Inglesi-Lotz [27] concluded that the service sector, which has a lower environmental impact, is expanding in more developed countries. Zhang Y. J. et al. [28] highlighted the increasing promotion of the service industry in national economies in recent years. Furthermore, Li R. et al. [29] concluded that the growing employment in the service sector has a positive environmental impact.

Servitization also exists in the energy sector. Park [22] pointed out the servitization model called “energy as a service” (EaaS). This model allows for the receipt of energy services without investing capital. He claimed that the spread of servitization in the energy sector can improve issues such as energy supply, trading, storage, delivery, and consumption, which are important aspects of Sustainable Development (see goal no. 7 of SD). The link between the servitization process and sustainability has been discussed in many studies [30,31,32]. It is also important to analyze sustainability (particularly energy sustainability) using a spatial approach. Broto and Baker [33] demonstrated the significance of space in energy studies in the context of technology diffusion, infrastructure development, political linkages, etc. In turn, Radmehr et al. [34] discussed the spatial spillovers of renewable energy on ecological sustainability. Gao et al. [35] presented a spatial approach to energy analysis, specifically the availability of electricity. Additionally, Osman et al. [36] and Zhang J. et al. [37] considered the interactions between various Sustainable Development Goals (SDGs) in Africa and China, respectively. The difference between these spatial studies is methodological. The first study is based on Local Moran statistics and cluster analysis. The authors concluded that there are clusters of high values for most SDGs in Northern Africa, as well as groups of countries with low values for almost all SDGs in the central part of the continent. The second study only focuses on the spatial distribution of the RV coefficient (which indicates the similarity between two matrices) in the context of the SDGs across the Chinese provinces. The authors compared the similarity of the SDG values in three groups: essential needs (natural science and technology), governance (interdisciplinary and policy), and objectives (social sciences and ethics).

The service-oriented production of individual manufacturers has a significant impact on the servitization of the global economy. In this research, the servitization process is discussed on a macroeconomic scale and measured by the share of value added by the services in the Gross Domestic Product (GDP).

Our approach to exploring the relationship between the process of economic servitization and energy sustainability is original, mainly for two reasons. Firstly, the impact of servitization on Sustainable Development, particularly energy sustainability, has rarely been empirically addressed on a macro scale in the literature. Secondly, we extend the spatial approach using the spatial econometric models incorporating economic similarity matrices to identify the spatial connections between countries, important for the relationship under consideration. The consideration of economic similarity is used to show the strength of the connections between the countries with the closest economic development in the case of energy sustainability. This details the general division of countries into developed and developing ones. The motivation for using this approach was that not all countries within the two presented regimes are closely related. Moreover, the transfer of the technology solutions generating the expansion of the services sector between the most similar states seems to be stronger than that between the more economically distant countries. It is also important to show the differences in the detected relationships when comparing the developed and developing countries.

3. Methods

The evaluation of the energy sustainability levels of the chosen world countries constitutes the first part of this research. Energy sustainability is identified with achieving the seventh goal of Sustainable Development. This is a process described by several diagnostic variables; therefore, the synthetic measure to its description is used. Our study used a slightly modified Perkal measure [38]. This is the composite indicator that requires several characteristics for describing the considered process. The diagnostic variables are composed into one measure characterizing the investigated phenomenon. The assessment of the level of the considered process with one value is helpful for the analysis of more complicated economic phenomena, such as energy sustainability, for example. The classical approach to calculate the values of the synthetic variables using this tool contains two stages [39].

i.

Transforming diagnostic variables into stimulants (as needed) and standardization of variable values to enable comparability:

$$x_{ij,t}^{\prime }={\displaystyle \frac{x_{ij,t}-{\overline{x}}_j}{S\left(x_j\right)}},$$

(1)

where $x_{ij,t}^{\prime }$ expresses the normalized value of the variable, $x_{ji,t}$ is the value of the $j$-th variable in the $i$-th country in time $t$, ${\overline{x}}_j$ denotes the arithmetic average of the $j$-th variable, and $S\left(x_j\right)$ is its standard deviation. The purpose of this step is to avoid the influence of differences resulting from the measurement units of the features.

ii.

Calculation of arithmetic average of standardized values:

$${PM}_{i,t}={\displaystyle \frac{\sum _{j=1}^p{x}_{ij,t}^{\prime }}{p}},$$

(2)

where ${PM}_{i,t}$ is the Perkal measure in the $i$-th country in time $t$, $p$ denotes the number of diagnostic variables, and $x_{ij,t}^{\prime }$—as above. For the needs of this research, in the third step, the Perkal measure is corrected with the $\delta$ value:

$${ES}_{i,t}={PM}_{i,t}+\delta ,$$

(3)

where ${ES}_{i,t}$ is the corrected Perkal measure of energy sustainability in the $i$-th country in time $t$, and $\delta =-\underset{i}{\min }{PM}_{i,t}+{\displaystyle \frac{1}{5}}S({PM}_{i,t})$, where $S\left({PM}_{i,t}\right)$ denotes the standard deviation of the Perkal measure. This transformation generates all values of $ES$ greater than zero that are necessary to estimate the sensitivity model in the next step of the investigation. Additionally, the $ES$ values are normalized by dividing by their maximums to get values between zero and one. Higher values of the $ES$ measure indicate better levels of the process considered.

The dependence of energy sustainability on economic structural changes is verified based on the spatio-temporal sensitivity model. The initial formula of the model is as follows [40]:

$$ln\left({ES}_{i,t}\right)={\xi }_{1}ln\left({SVA}_{i,t}\right)+{\sum }_{k=0}^p{\sum }_{m=0}^p{\sum }_{l=0}^p{\theta }_{kml}{x}_i^k{y}_i^m{t}^l+{\epsilon }_{i,t},$$

(4)

where $ln\left(ES\right)$ and $ln\left({SVA}_{i,t}\right)$ denote the natural logarithms of the energy sustainability measure and the share of value added by services in the GDP, respectively. In turn, $\sum _{k=0}^p\sum _{m=0}^p\sum _{l=0}^p{\theta }_{kml}{x}_i^k{y}_i^m{t}^l$ is the spatio-temporal trend component [41,42,43], whereas $x_i$ and $y_i$ are the geographical location coordinates of the $i$-th country. The parameter ${\xi }_1$ is called the sensitivity parameter. It measures the percentage change in $ES$ value caused by the 1% change in service value-added.

The model (4) is then enriched with the spatial dependence between the neighboring territorial units. The occurrence and character of the spatial dependence are discussed with the Moran $I$ statistics [44] and Lagrange Multiplier tests [45], respectively. As a result, the spatial autoregressive (SAR) and spatial error (SE) models take the following forms:

$$ln\left({ES}_{i,t}\right)={\xi }_{1}ln\left({SVA}_{i,t}\right)+{\sum }_{k=0}^p{\sum }_{m=0}^p{\sum }_{l=0}^p{\theta }_{kml}{x}_i^k{y}_i^m{t}^l+\rho {\mathit{W}}_{\mathit{t}}ln\left({SVA}_{i,t}\right)+{\epsilon }_{i,t},$$

(5)

$$\begin{gathered} ln\left({ES}_{i,t}\right)={\xi }_{1}ln\left({SVA}_{i,t}\right)+{\sum }_{k=0}^p{\sum }_{m=0}^p{\sum }_{l=0}^p{\theta }_{kml}{x}_i^k{y}_i^m{t}^l+{\epsilon }_{i,t}, \\ {\eta }_{i,t}=\lambda {\mathit{W}}_{\mathit{t}}{\eta }_{i,t}+{\epsilon }_{i,t}, \end{gathered}$$

(6)

where ${\mathit{W}}_{\mathit{t}}ln\left({SVA}_{i,t}\right)$ denotes the average value of the SVA from the neighboring countries and ${\mathit{W}}_{\mathit{t}}$ is the constant in time of a spatial connections matrix, and the spatial dependence parameters $\rho$ and $\lambda$ are estimated and verified.

This research defines the neighborhood in two ways. The first of them quantifies the neighborhood by the land common border criterion (the matrix $\mathit{W}$ is used). It is the classical and most often used spatial weights matrix in the spatial analyses. Secondly, the spatial connection matrix was constructed on the basis of the economic distance between units (marked by $\mathit{D}$). The variable characterizing the economic similarity of the spatial units is the Human Development Index (HDI). The choice of the HDI allows us to check whether the division into the regimes of developed and developing countries in the next step of the research is valid, in particular. The distance matrix is built in five steps [46,47]: (1) the calculation of the distance between the pairs of countries using the absolute value of the differences in HDI, (2) the determination of the borderline level of similarity between the countries, (3) the change of the distance matrix values above the borderline level to zero, (4) the inversion of the nonzero elements of the distance matrix, and (5) the standardization of the matrix values by rows to one.

The last part of the research contains an analysis of the considered relationship in two regimes, i.e., the developed and developing countries. For the analysis in the two separate regimes, models (4), (5), and (6) take the forms given as Equations (7), (8), and (9), respectively:

$$ln\left({ES}_{i,t}\right)={\sum }_h{\xi }_1^h{\left({SVA}_{i,t}\right)}^h+{\sum }_{k=0}^p{\sum }_{m=0}^p{\sum }_{l=0}^p{\theta }_{kml}{x}_i^k{y}_i^m{t}^l+{\epsilon }_{i,t},$$

(7)

$$ln\left({ES}_{i,t}\right)=\sum _h{\xi }_1^h{\left({SVA}_{i,t}\right)}^h+{\sum }_{k=0}^p{\sum }_{m=0}^p{\sum }_{l=0}^p{\theta }_{kml}{x}_i^k{y}_i^m{t}^l+\rho {\mathit{W}}_{\mathit{t}}ln\left({SVA}_{i,t}\right)+{\epsilon }_{i,t},$$

(8)

$$\begin{gathered} ln\left({ES}_{i,t}\right)=\sum _h{\xi }_1^h{\left({SVA}_{i,t}\right)}^h+{\sum }_{k=0}^p{\sum }_{m=0}^p{\sum }_{l=0}^p{\theta }_{kml}{x}_i^k{y}_i^m{t}^l+{\eta }_{i,t}, \\ {\eta }_{i,t}=\lambda {\mathit{W}}_{\mathit{t}}{\eta }_{i,t}+{\epsilon }_{i,t}, \end{gathered}$$

(9)

where h ∈ {1,2} is the considered regime.

The proposed approach has certain limitations. Firstly, the composite indicator approach used in this research does not differentiate the importance of the defined determinants of energy sustainability. All diagnostic variables are treated as equally important. Moreover, the spatial range of the study does not allow us to choose many appropriate characteristics of energy sustainability. This is the eternal choice of spatial research: more territorial units in the investigation or a wider description of the considered process. Therefore, one of the limitations is the lack of the division of energy into groups of specific types, as in the study presented by Nassar et al. [48].

Moreover, the investigation presents some uncertainties. Above all, the data from some countries may be burdened with a large measurement error, particularly in the developing countries (Africa, South Asia, etc.). The choice of the HDI as the best measure of the economic development similarity among countries to construct the spatial weights matrix can be debatable. However, the discussion about the choice of the economic distance matrix will be the subject of further research.

4. Data and Empirical Analysis

4.1. Energy Sustainability Assessment

The first part of the study concerns the assessment of energy sustainability. Because this process is multi-dimensional, several characteristics describe it.

Table 1. Diagnostic variables of energy sustainability.

Symbol	Variable	Character
$X_1$	Proportion of population with access to electricity (%)	S
$X_2$	Energy intensity level of primary energy (MJ/$2017 PPP GDP)	D
$X_3$	Renewable energy share of total final energy consumption (%)	S

presents a list of the diagnostic variables chosen to represent energy sustainability. The analysis is conducted for 164 selected countries worldwide (56 developed and 108 developing) from 2015 to 2020, which is why the list of variables is short. It is an eternal dilemma in taxonomic analyses: what is better, characterizing the process with as many indicators as possible or comparing as many territorial units as possible? The second approach is adopted in this research. The data describing energy sustainability come from the SDG Global Database run by the United Nations (https://unstats.un.org/sdgs/dataportal, accessed on 15 January 2024).

All diagnostic variables concern the achievement of the seventh goal of SD, called “ensuring access to affordable, reliable, sustainable and modern energy” (United Nations, 2015). The first variable characterizes the share of the population with access to electricity, and the second is the energy intensity level of the primary energy, defined as the energy supplied to the economy per unit value of economic output. In turn, the third variable refers to the proportion of renewable energy use. The first and third variables have a positive impact on energy sustainability, which is why they are marked as stimulants (S), whereas the second one has an opposite impact to the stimulants and is marked as D (see Table 1).

Table 2. Results of energy sustainability assessment—values of ES measure and rankings.

Country	ES 2015	ES 2020	Rank 2015	Rank 2020
Puerto Rico	0.3319	0.3326	1	1
Sri Lanka	0.2315	0.2369	2	3
Bhutan	0.2297	0.2373	3	2
Uruguay	0.2274	0.2277	4	5
Paraguay	0.2252	0.2237	5	7
Iceland	0.2207	0.2268	6	6
Nepal	0.2196	0.2157	7	15
Gabon	0.2177	0.2328	8	4
Norway	0.2156	0.2235	9	8
Costa Rica	0.2112	0.2153	10	17
Guatemala	0.2098	0.2186	11	11
Sweden	0.2098	0.2186	12	12
…
Vietnam	0.1818	0.1705	44	88
…
Saint Lucia	0.1692	0.1646	71	113
…
Mali	0.1671	0.1612	75	117
…
Bahamas	0.1666	0.1601	80	120
…
Tonga	0.1628	0.1516	93	133
…
Venezuela	0.1619	0.1995	97	27
…
Ethiopia	0.1615	0.1865	99	50
…
Kenya	0.1603	0.1950	101	36
…
Rwanda	0.1594	0.1827	107	57
…
Afghanistan	0.1544	0.1801	120	64
…
Zimbabwe	0.1513	0.1761	123	74
…
Uganda	0.1485	0.1749	131	77
…
Russia	0.1363	0.1412	153	147
Namibia	0.1346	0.1357	154	154
Chad	0.1324	0.1342	155	155
Mongolia	0.1324	0.1437	156	145
Turkmenistan	0.1324	0.1339	157	157
South Africa	0.1317	0.1392	158	150
Trinidad and Tobago	0.1307	0.1312	159	159
Syria	0.1236	0.1234	160	162
Mauritania	0.1228	0.1204	161	163
Benin	0.1212	0.1295	162	160
Papua New Guinea	0.1158	0.1169	163	164
Lesotho	0.1129	0.1265	164	161

presents selected values of the energy sustainability ($ES$) measure in 2015 and 2020, as well as the rankings generated based on them. In both years considered, Puerto Rico held the first place, despite its relatively low share of renewable energy in total energy consumption. Throughout the entire analyzed period, Puerto Rico had a very low energy intensity, and the entire population had access to electricity. The situation was slightly different in Bhutan, which ranked third in 2015. There, the variable $X_2$ had one of the highest values, but almost all energy consumed came from renewables (a very high value of the variable $X_3$).

It is worth noting that most of the Scandinavian countries—Iceland, Norway, and Sweden—were among the top twelve countries. In Norway and Sweden, the variable $X_2$ was at a relatively low level, and around half of the consumed energy came from renewables. On the other hand, Iceland had a higher level of energy intensity and a higher share of renewable energy use.

At the bottom of the rankings were countries such as Syria, Mauritania, Benin, Papua New Guinea, and Lesotho. In the case of Syria, the position in the hierarchy was influenced by very high values and very low values of the variables $X_2$ and $X_3$, respectively. Mauritania, Benin, and Papua New Guinea had a lower proportion of the population with access to electricity. In Lesotho, in addition to the low level of the variable $X_1$, the energy intensity was also high. There were no significant changes in these countries’ positions at the top and bottom of the rankings between 2015 and 2020.

The greatest improvements in the ranking were observed in the case of Venezuela (up 70 places), Kenya (up 65 places), Afghanistan (up 56 places), Uganda (up 54 places), Rwanda (up 50 places), Ethiopia, and Zimbabwe (both up 49 places). In Venezuela, there was a decrease in the energy intensity from 2015 to 2020. In the remaining countries, the improvement was caused by an increase in the population with access to electricity. On the other hand, the most negative changes in the hierarchy were seen for Vietnam (down 44 places), Saint Lucia, Mali (both down 42 places), Bahamas, and Tonga (both down 40 places). In Vietnam, Saint Lucia, and Mali, there was a decrease in the consumption of energy from renewables between 2015 and 2020. The fall in ranking for the Bahamas and Tonga was due to a lack of improvement in the levels of the variables $X_2$ and $X_3$.

Figure 1 presents the spatial differentiation of the energy sustainability in the extreme years of the study. Part (a) of the figure is for 2015, while part (b) refers to 2020. In both years, a certain trend was observed. Most countries in Latin and South America, as well as Europe, had values of energy sustainability above the median. On the other hand, most countries with the lowest values of the ES measures were located in Africa and southwestern Asia. Furthermore, Africa and Asia had the highest differentiation of energy sustainability between states.

The observations made based on the maps in Figure 1 are confirmed by the spatio-temporal trend model, the results of the estimation and verification of which are shown in

Table 3. The results of the spatio-temporal structure analysis of the energy sustainability assessment.

Parameter	Estimate	Std. Error	t-Statistics	p-Value
${\mathit{\theta }}_{000}$	−1.8030	0.0114	−158.6110	0.0000
${\mathit{\theta }}_{100}$	−0.0005	0.0001	−6.3710	0.0000
${\mathit{\theta }}_{010}$	0.0008	0.0002	4.1970	0.0000
${\mathit{\theta }}_{001}$	0.0067	0.0028	2.4080	0.0162
Moran’s $I$ (p-value)
$\mathit{W}$ matrix	0.1919 (0.0000)
$\mathit{D}$ matrix	0.0839 (0.0000)

. The estimates of the parameters ${\theta }_{100}$ and ${\theta }_{010}$ (negative and positive, respectively) indicate that the energy sustainability decreased on average from west to east and from north to south. Additionally, an average increase in the synthetic measure values over time was observed (see parameter ${\theta }_{001}$).

The results of the Moran test show that there is spatial dependence between countries in relation to energy sustainability. Both the geographical ($\mathit{W}$ matrix) and economic ($\mathit{D}$ matrix) neighborhoods turned out to be statistically significant, but the connections between the countries adjacent in space are stronger. The importance of the similarity in energy sustainability provides a basis for conducting further analysis by dividing the spatial units considered into regimes of developed and developing countries.

4.2. Assessment of Sensitivity of Energy Sustainability on Servitization Process

The second part of the study starts with the assessment of the spatio-temporal structure of the service value-added in GDP. This stage allows for evaluating the global tendencies occurring in the servitization process. The data characterizing the servitization process were downloaded directly from the World Bank Indicators database (https://data.worldbank.org/indicator, accessed on 15 January 2024).

Figure 2 presents the spatial differentiation of the share of the service value-added in the Gross Domestic Product in the following years: 2015—part (a) and 2020—part (b). Observing the maps, it is worth noting that the highest values of the considered variable in 2015 and 2020 were in relatively highly developed countries, such as the USA, Canada, countries located in the northern and western parts of Europe, and Australia. Almost all states that belonged to the group with the lowest values were located in Africa. This indicates that the less developed countries focus on industrial development, which allows faster economic growth for them.

When analyzing the spatio-temporal structure of the considered process, it should be noted that the direction of the changes in this case was the same as in the energy sustainability. The share of service value-added in the GDP, on average, decreased in the eastern and southern directions. Therefore, it can be presumed that the servitization process favors energy sustainability. However, the time variable in the spatio-temporal trend model is not significant (see the estimate of ${\theta }_{001}$ in

Table 4. The results of the spatio-temporal structure assessment of the servitization process.

Parameter	Estimate	Std. Error	t-Statistics	p-Value
${\mathit{\theta }}_{000}$	3.9870	0.0144	276.1330	0.0000
${\mathit{\theta }}_{100}$	−0.0006	0.0001	−7.1620	0.0000
${\mathit{\theta }}_{010}$	0.0014	0.0002	5.8060	0.0000
${\mathit{\theta }}_{001}$	0.0004	0.0035	0.1120	0.9110
Moran’s $I$ (p-value)
$\mathit{W}$ matrix	0.2862 (0.0000)
$\mathit{D}$ matrix	0.3140 (0.0000)

). This is undoubtedly the result of a different growth rate of servitization across the selected countries. The servitization process was observed in almost all developed countries (53 out of 56), but in less than half of the developing ones (42 out of 108). As a result, the average world servitization process is not statistically significant.

Moreover, in the case of the servitization process, the economic similarity of the countries is slightly more important than their physical neighborhood. The Moran $\mathit{I}$ statistic for the $\mathit{D}$ matrix is higher than for the $\mathit{W}$ matrix.

Table 5. The results of the estimation and verification of the spatio-temporal sensitivity model of energy sustainability on the servitization process.

Parameter	Estimate	Std. Error	t-Statistics	p-Value
${\mathit{\theta }}_{000}$	−2.3910	0.0992	−24.1030	0.0000
${\mathit{\theta }}_{100}$	−0.0004	0.0001	−4.9900	0.0000
${\mathit{\theta }}_{010}$	0.0006	0.0002	3.1110	0.0019
${\mathit{\theta }}_{001}$	0.0066	0.0027	2.4290	0.0153
${\mathit{\xi }}_1$	0.1474	0.02472	5.962	0.0000
Matrix:	$\mathit{W}$		$\mathit{D}$
Moran’s $I$	0.1819 (0.0000)		0.0418 (0.0056)
LM tests
${\mathit{L}\mathit{M}}_{\mathit{S}\mathit{E}}$	62.0968 (0.0000)		6.0842 (0.0136)
${\mathit{L}\mathit{M}}_{\mathit{S}\mathit{A}\mathit{R}}$	2.1805 (0.1398)		15.7181 (0.0001)
${\mathit{R}\mathit{L}\mathit{M}}_{\mathit{S}\mathit{E}}$	59.9501 (0.0000)		26.1854 (0.0000)
${\mathit{R}\mathit{L}\mathit{M}}_{\mathit{S}\mathit{A}\mathit{R}}$	0.0339 (0.8540)		35.8193 (0.0000)

shows the results of the estimation and verification of the spatio-temporal sensitivity model in the relationship between energy sustainability and the servitization process. The most important parameter, ${\xi }_1$, is positive and statistically significant. Its value means that a one percent increase in the share of service value-added in the GDP causes a growth in the energy sustainability measure by 0.1474 percent. This indicates that the progressive servitization process favors energy sustainability. The Moran test results indicate the occurrence of spatial autocorrelation in the model residuals for both connection matrices, $\mathit{W}$ and $\mathit{D}$, whereas the spatial dependence for the first matrix is stronger.

The Lagrange Multiplier tests for the geographical neighborhoods show significance only for the spatial error (SE) model. The p-values over 0.05 correspond to statistics ${LM}_{SAR}$ and ${RLM}_{SAR}$, which are responsible for the choice of the spatial autoregressive model. The situation for the economic proximity matrix is different. There, all LM statistics are statistically significant, but ${LM}_{SAR}$ and ${RLM}_{SAR}$ have higher values than ${LM}_{SE}$ and ${RLM}_{SE}$, respectively. As a result, the spatial autoregressive model is considered better. It is worth noting that when the economic neighborhood is used, the model with a more valuable interpretation is adopted. This is an important issue to prefer economic proximity over geographical proximity.

Table 6. The estimation and verification results of the SAR and SE models of the sensitivity of energy sustainability on the servitization process.

Model	$\mathit{W}$_SE		$\mathit{D}$_SAR
Parameter	Estimate	p-Value	Estimate	p-Value
${\mathit{\theta }}_{000}$	−2.3239	0.0000	−1.8452	0.0000
${\mathit{\theta }}_{100}$	−0.0003	0.0001	−0.0004	0.0000
${\mathit{\theta }}_{010}$	0.0007	0.0062	0.0004	0.0213
${\mathit{\theta }}_{001}$	0.0058	0.1021	0.0050	0.0635
${\mathit{\xi }}_1$	0.1323	0.0000	0.1152	0.0000
$\mathit{\rho }$	-	-	0.2308	0.0001
$\mathit{\lambda }$	0.3605	0.0000	-	-
Moran’s $I$	0.0147 (0.2922)		−0.0134 (0.2312)

presents the results of the estimation and verification of the sensitivity models with spatial dependence. The spatial error model is estimated for matrix $\mathit{W}$, while the spatial autoregressive model is estimated for matrix $\mathit{D}$. The parameter ${\xi }_1$ remains statistically significant, but its value is slightly lower than in the base model: 0.1323 and 0.1152 for the SE and SAR model, respectively. The sensitivity of energy sustainability on the servitization weakens when the spatial connections between countries are introduced in the model.

More importantly, the spatial dependence parameters $\lambda$ and $\rho$ are also statistically significant. The estimate of the parameter $\lambda$ indicates that variables omitted in the model or random components from the neighboring countries influence the energy sustainability in a given country. A positive estimate of the parameter $\rho$ denotes a similarity in the levels of the energy sustainability measure in countries with a similar level of the Human Development Index. Hence, in the next part of the study, the analysis is provided with the division of countries into two separate regimes: developed (the first regime) and developing (the second regime).

Table 7. The estimation and verification results of the spatio-temporal model of the sensitivity of energy sustainability on the servitization process in the regimes.

Parameter	Estimate	Std. Error	t-Statistics	p-Value
${\mathit{\theta }}_{100}$	−0.0004	0.0001	−5.1580	0.0000
${\mathit{\theta }}_{010}$	0.0005	0.0002	2.2520	0.0246
${\mathit{\theta }}_{001}$	0.0064	0.0027	2.3800	0.0175
${\mathit{\theta }}_{000}^1$	−2.9630	0.2805	−10.5650	0.0000
${\mathit{\xi }}_1^1$	0.2886	0.0677	4.264	0.0000
${\mathit{\theta }}_{000}^2$	−2.1930	0.1212	−18.0940	0.0000
${\mathit{\xi }}_1^2$	0.0968	0.0306	3.1680	0.0016
Matrix:	$\mathit{W}$		$\mathit{D}$
Moran’s $I$	0.1750 (0.0000)		0.0562 (0.0003)
LM tests
${\mathit{L}\mathit{M}}_{\mathit{S}\mathit{E}}$	57.4642 (0.0000)		10.9930 (0.0009)
${\mathit{L}\mathit{M}}_{\mathit{S}\mathit{A}\mathit{R}}$	2.2992 (0.1294)		17.9610 (0.0000)
${\mathit{R}\mathit{L}\mathit{M}}_{\mathit{S}\mathit{E}}$	55.2374 (0.0000)		13.7400 (0.0002)
${\mathit{R}\mathit{L}\mathit{M}}_{\mathit{S}\mathit{A}\mathit{R}}$	0.0724 (0.7878)		20.7080 (0.0000)

shows the results of the estimation and verification of the spatio-temporal sensitivity model in the different regimes. The most important aspect is the interpretation of the properties of parameters ${\xi }_1^1$ and ${\xi }_1^2$, which provide information about the impact of the servitization process on the energy sustainability in developed and developing countries. Both of these parameters are statistically significant, but the estimate of parameter ${\xi }_1^2$ is much lower than that of ${\xi }_1^1$. This indicates that the servitization process has had a weaker impact on energy sustainability in the developing countries during the period of 2015–2020. The reason for this difference is the more advanced progress of the servitization process in almost all developed countries, compared to less than half of the developing countries. The developing countries instead focus on other elements of their economy that contribute to faster economic growth at their current level of development.

Similar to the model without regimes, there is spatial autocorrelation in the residuals for both connection matrices. In addition, the geographical proximity shows a stronger correlation. Based on the results of the LM tests, the SE model is chosen for the geographical neighborhood, while the SAR model is chosen for the economic neighborhood. The results of the estimation and verification of the selected models are presented in

Table 8. The estimation and verification results of the SAR and SE models of the sensitivity of energy sustainability on the servitization process in the regimes.

Model	$\mathit{W}$_SE		$\mathit{D}$_SAR
Parameter	Estimate	p-Value	Estimate	p-Value
${\mathit{\theta }}_{100}$	−0.0003	0.0001	−0.0004	0.0000
${\mathit{\theta }}_{010}$	0.0006	0.0163	0.0003	0.1121
${\mathit{\theta }}_{001}$	0.0057	0.1025	0.0047	0.0791
${\mathit{\theta }}_{000}^1$	−2.7906	0.0000	−2.5235	0.0000
${\mathit{\xi }}_1^1$	0.2460	0.0003	0.2884	0.0000
${\mathit{\theta }}_{000}^2$	−2.2217	0.0000	−1.6040	0.0000
${\mathit{\xi }}_1^2$	0.1062	0.0005	0.0610	0.0425
$\mathit{\rho }$	-	-	0.2467	0.0000
$\mathit{\lambda }$	0.36497	0	-	-
Moran’s $I$	0.0119 (0.3254)		−0.0034 (0.4440)

.

Adding spatial dependence to the sensitivity models with the regimes, parameters ${\xi }_1^1$ and ${\xi }_1^2$ are still statistically significant. Nevertheless, differences in their values are observed. Firstly, parameter $\lambda$ caused a decrease and an increase in parameters ${\xi }_1^1$ (from 0.2886 to 0.2460) and ${\xi }_1^2$ (from 0.0968 to 0.1062), respectively. Therefore, the impact of omitted variables or random processes from the neighboring countries weakened and strengthened the sensitivity of the energy sustainability on the servitization process in the first and second regimes, respectively.

Considering the economic similarity, the value of parameter ${\xi }_1^1$ did not change significantly, but the value of parameter ${\xi }_1^2$ is much lower than in the basic model. This indicates a decrease in sensitivity in the considered relationship in the group of developing countries when including the similarity of the energy sustainability in the countries with similar HDI levels. Additionally, both models do not show spatial autocorrelation.

5. Conclusions

Changes in the structure of the economy towards servitization are observed in most countries worldwide. The intensity of this process is connected to the level of development. The growth in the share of service value-added in the Gross Domestic Product is most evident in developed countries. Some developing countries, aiming for faster economic growth, particularly focus on developing their industrial sector. Others, due to their location and geographical characteristics, are more inclined to focus on agricultural production. However, many of them are transforming their production by introducing services that are more environmentally friendly than the industrial sector. Therefore, more developing countries are showing a move towards a service-based economy.

Many researchers have concluded that a focus on servitization by manufacturers has a positive impact on the natural environment and, as a result, contributes to sustainability. This research, unlike most existing studies, focuses on the servitization process in macroeconomic terms and its influence on energy sustainability. Energy sustainability was chosen as a particular topic of Sustainable Development due to the crucial impact of the changes in the energy sector on the natural environment. Evaluating the level of energy sustainability, based on the seventh goal of Sustainable Development, African and southwestern Asian countries (recognized as developing ones) were found to be at the bottom of the rankings during the extreme years of the investigation. Generally, the developed states showed a higher level of energy sustainability between 2015 and 2020 than the developing ones, mainly due to a higher share of the population having access to electricity. The issue of the availability of amenities, mostly in African countries, is well known.

The analysis of the relationship between the progress of the servitization process and the energy sustainability indicates that a focus on the service sector significantly improves sustainability in the energy sector. Therefore, the changes in the economy towards servitization have a positive impact on Sustainable Development in this regard. However, this relationship also depends on the level of development of the countries. This is evident in the second part of the analysis, where the energy sustainability was found to be more sensitive to the servitization process. In the case of the developing states, almost all showed growth in the energy sector from 2015 to 2020, but an increase in the share of service value-added in the GDP was observed in only around 40 percent of them.

Additionally, the geographical location and economic development similarities between the countries significantly influence the structure of the economy and energy sustainability worldwide. These similarities were included in the study, using tools such as spatial connection matrices based on common borders and the similarity in the Human Development Index (HDI) levels, respectively, in the spatial models. Despite the stronger dependence in the case of geographical proximity, the similarity in economic development provided a more valuable interpretation through the estimation of a spatial autoregressive model (SAR). Therefore, the development similarity better explains the relationship between the servitization process and the energy sustainability than the geographical location of the countries does.

This research provides several important insights for policymakers. Firstly, they can find one of the methods for ensuring energy sustainability. They should focus on introducing regulations that provide for the expansion of the service sector. Governments from less developed countries should be inspired by the rulers from the relatively more developed ones. Above all, through international and national policies, relatively poor countries should support the expansion of the service sector to reduce emissions and costs and increase efficiencies, as is done in developed economies. Moreover, policymakers should focus on the improvement of the individual determinants of energy sustainability, such as renewable energy use. Hence, it is required that the introduced regulations encourage investment in renewable energy sources.

The use of renewable energy and the servitization of the economy increase sustainable practices. One of the results is the reduction in greenhouse gas emissions per unit of energy consumption. Therefore, economies (both developed and developing) should strive to shift production towards the use of services.

In further research, it is worth discussing the influence of the servitization process on the other aspects of Sustainable Development. A three-pillar approach to Sustainable Development could be used. In the future, we will analyze the relationship between Sustainable Development and the expansion of other sectors of the economy.