Does Energy Demand Security Affect International Competitiveness? Case of Selected Energy-Exporting OECD Countries

Abstract

International competitiveness and energy security are important topics on the energy policy agenda of energy-exporting and -importing nations. High dependence on energy rents challenges exporters’ economies and influences their ability to compete on international markets. The goal of this study is to investigate how energy demand security affects the international competitiveness of developed energy exporters. This research employs an econometric approach aimed at modeling the Heckscher–Ohlin and Ricardo international trade hypotheses. Introduced modifications allow for the measurement of international competitiveness on a country level, not per industry, and includes a series of energy variables in addition to relative resource endowment and productivity differentials. This study (based on Norway, the Netherlands, Canada and the USA over the period of 1997–2017) proves that energy security does not play a role in the shaping of international competitiveness of high-income energy-exporting countries.

1. Introduction

International competitiveness is one of the most important topics on the policy agenda; so is energy security. These two concepts are at the heart of current economic and energy debates. Every nation, regardless of its economic standing, places great emphasis on improving its ability to compete on international markets, with energy security being of similar importance. Stable energy supplies determine continuous production and consumption. Affordable energy prices affect industrial competitiveness and improve access to modern energy services.

Even though both phenomena are vital for every nation, they influence countries differently. Highly internationally competitive nations will be rather focused on moving their competitive edge towards innovations and R&D-driven investments, whereas countries struggling with basic systematic challenges will be more interested in efficiency-supported solutions. Similar gradual development is visible in terms of energy security. Countries with unsatisfied basic energy needs will be more inclined to meet the energy demand with available and affordable energy resources, while economies with full electrification rates and stable access to modern energy services will be more interested in fitting their energy consumption patterns to climate/environmental challenges, as presented in the energy needs pyramid of Frei [1]. Similarly, a link between energy security and international competitiveness will be different for different nations, depending on their position in global energy markets. Countries exporting energy will look at energy security from the point of view of the stability of export incomes and their influence on international competitiveness [2,3]. On the other hand, countries importing energy commodities will pay special attention to the determinants of energy prices and the impact they will have on the competitiveness of domestic industries, including those producing export [4,5,6,7,8,9].

The importance of energy demand security for the international competitiveness of exporting nations was proven by the Dutch disease phenomenon in the Netherlands in the 20th century. The Dutch case demonstrated how oil and gas export revenues affected international competitiveness of an economy, its manufacturing sector, domestic currency exchange rates and wages. In the 21st century, exporting nations, both developed and developing, learned their lessons from the Dutch disease case. They insulated their economies against disease symptoms by introducing fiscal policies and financial mechanisms allocating excess of export revenues. Through sovereign wealth funds, developed and developing countries managed excess in oil and gas export incomes with responsibility towards next generations. Nowadays, it is not as important how to manage windfall profits, but rather how to stay competitive on international markets in non-energy industries. For developing countries, that question is irrelevant, as their economies are mostly dependent on export of energy resources. However, for advanced nations with more diverse trade structures, international competitiveness of economies and non-energy-exporting industries is of high importance. Yet the problem of managing oil incomes has been accompanying Norway’s energy policy since oil and gas fields under the North Sea were discovered in the 20th century—as confirmed by the Central Bureau of Statistics for Norway [10] already in 1982. More recent studies [5,6] of the International Monetary Fund (IMF) highlighted that problem arguing that “Norway’s half century of good fortune from its oil and gas wealth may have peaked. Oil and gas production will continue for many decades on current projections, but output and investment have flattened out, and the spillovers from the offshore oil and gas production to the mainland economy may have turned from positive to negative” [5]. In 2015, the IMF [5] predicted that the Norwegian mainland economy was likely to be negatively affected by sustained changes in oil prices. A few years later, the IMF [6] stressed that competitiveness of the non-oil manufacturing sector is a challenge for Norway. The Norwegian case shows that even high-income countries may face challenges with energy rents’ influence over international competitiveness. In the future, the importance of energy demand security challenges in Norway will rise, especially once the Green Deal will be fully introduced in the European Union (EU). Fischer [11] stressed the significance of the Green Deal for Norway’s future as an energy exporter and argues that as the “EU’s imports of gas will decline over time, a smaller market will be more strongly contested. Thus, Norway will face energy security challenges” (understood as demand security).

The aim of this study is to investigate the influence of energy security on international competitiveness of developed, energy-exporting economies. To do so, econometric modeling is used as a tool for testing the research hypothesis: relative energy demand security has a positive and a statistically significant impact on relative international competitiveness. This study fits into the research gap highlighted by Nyga-Łukaszewska and Chilimoniuk-Przeździecka [7] calling for broader empirical investigation of the energy demand security-international competitiveness link, especially in the group of countries divided upon their GDP performance. Additionally, the research niche, that is tested in this paper, corresponds to an appeal of Antoniadis [12], who recommends extending the oil rents-international competitiveness discussion to other than Black Sea energy-exporting economies. Apart from that, sole matter of energy security in resource rich exporting nations (other than Russia and the Caspian region), is indicated as a research gap by Karatayev and Hall [13]. This paper contributes to the existing scientific knowledge by looking at the energy demand security influence (also depicted through oil rents) on the international competitiveness of high-income OECD energy exporters. We decided to choose these countries as they exhibit comparable level of economic development and have similar standards in terms of quality of institutions (rule of law, bureaucracy, corruption etc.). In this way our sample is coherent and does not include any outliers.

This research departs from the existing literature in various ways. Firstly, study employs the concept of demand security, whereas most studies focus on supply security–as shown by Dike [8]. Secondly, the analysis on demand security has so far been carried out for OPEC countries [9], Saudi Arabia [14] or Russia [15]. Studies looking at the industrialized nations, among them the USA (or the US), Australia, Canada, Norway and the Netherlands, mainly focused on supply security perspective [16,17,18,19,20]. Thirdly, this study embeds the energy security-international competitiveness analysis within a wider framework of international trade models (testing Ricardo and Heckscher–Ohlin trade approaches). Obtained results show that relative energy demand security does not have a statistically significant impact on the relative international competitiveness of developed, energy-exporting economies. Lastly, empirical results are stable across various approaches used (GLS, PCSE, FE) in the econometric modeling process.

The study is organized as follows. After the introduction, the theoretical framework is described. This part is followed by materials (data sources) and the empirical method, and a presentation of the results (econometric model with associated tests). The study ends with discussion and conclusion sections. In this text, a notion “energy demand security” is interchangeably used with a phrase a “energy security of exporters”, “security of energy-exporting nations” or “demand security”, while “energy supply security” is treated as the same as “energy security of importers”, “security of energy importing nations” or “supply security”. Similarly, the “developed nation” concept is synonymous for an advanced, high-income or industrialized country.

2. Theoretical Framework

This study uses the concept of energy security of energy-exporting nations, often called demand security [3] in contrast to supply security. The latter idea has been widely recognized in the literature, even though researchers admit that it is a complex [21,22,23], ubiquitous [24,25,26] and a blurred concept [19] without one common definition [27,28,29,30]. Only between the years 2000 and 2014, around 100 papers had been presented with more than 80 different energy security definitions [31]. As the International Energy Agency (IEA) [32] suggested, “energy security is the uninterrupted availability of energy sources at an affordable price”. The IEA was established in the 20th century with as a core mission to ensure energy security to industrialized countries. With time, energy markets evolved and countries were confronted with new challenges. Energy security remains an important policy and an equally important research question [9], especially for energy-importing nations. However, recently, a new phenomenon has emerged: energy demand security. This is a relatively “young” idea, dating back to the beginnings of the 21st century. This concept highlights energy exporters perspective. Oftentimes, exporters’ perspective relies on the achievements of supply security studies (described in, e.g., Ayoo [33]), as time-wise supply research on the supply security appeared earlier in the theoretical studies. Recognizing its importance for demand security, our research also employs a similar perspective by looking at demand security through existing research on supply security.

Researchers focusing on demand security perspective stress that this concept is overlooked in the literature [34] and, in contrast to supply security, is not defined by international institutions [9]. Only Energy Charter Secretariat [35] has recognized the differences between these two ideas. Dike [8] compared energy demand security to energy demand shocks that producers experience. He described energy demand security within four dimensions, which strongly correspond with a country’s dependence on the energy export. The first of these aspects refers to energy export dependence. Dike [8] claimed that the larger the quantity of crude oil and gas exports in a country’s total export value, the higher the energy export’s demand risks. The second dimension pertains to the so-called monopsony factor. This approach is understandable, as Dike [8] analyzed the situation of OPEC countries. He argued that the larger the size of a country’s import share of OPEC members’ export, the higher the country’s monopsony power on OPEC member states. Similarly, with the increase in the monopsony power of any individual importing country, the higher the risk exposure of OPEC member states. The third aspect in Dike’s [8] definition depicts transactional costs. It is assumed that the greater the distance between exporting and importing countries is and the higher the number of transit countries, the higher the risk is of energy export demand. The last, fourth aspect is economic dependence. Dike [8] assumed that the more dependent the economy is on energy export, the higher the security of demand risks. All four above-mentioned aspects are treated as risks contributing to demand security index. The higher the index, the riskier the energy export demand security is.

Novikau [15,36], on the other hand, offered a more general approach to energy demand security. He argued that this idea includes “the availability of customers, price fairness, and energy transit security issues, such as control over pipelines”. Novikau [36] believes that only demand—not supply—security matters for energy exporters. Bolino et al. [9] provided a more detailed analysis looking at the demand security through two major categories of energy security threats—physical supply disruption threats and economic instability—that apply both to importers and exporters. Bolino et al. [9] added to that approach a set of challenges for energy exporters that stem from analysis of the official policy documents of OPEC, GECF, Russia, Iran, Australia and Canada. They argued that these challenges include: macroeconomic dynamics and slowing demand for energy (and specific fuels) in importing economies, global energy price fluctuations, increased competition resulting from the emergence of new exporters, tariff and non-tariff barriers in importing countries, sanctions imposed on energy exporters or any buyer-related risks (breach of contract or default) and shifts in legislation in importing economies. This approach strongly resembles Romanova [3], who depicted energy demand security within five different aspects. First (possibly the best known) is price stability, second is oil and gas consumption stability, third is the level of taxes imposed on fuels in importing nations, fourth is the stability of regulation on consumers’ markets and, finally, fifth is the degree of competition in export markets. Romanova [3] argued that demand security guarantees stable budget revenues for exporting nations and that this money is used to finance the development of producing countries. She maintained that general “stability of exporting countries hinges on stability of oil and gas money flows” [3]. Similarly, Dannreuther [34] highlighted that the essence of energy demand security is a stable and secure revenue for development. Thus, demand security strongly corresponds with natural resource rents.

While looking at international competitiveness of countries, it is acknowledged that there is no consensus regarding its definition [37]—putting aside general definitions (e.g., “a measure of a country’s advantage or disadvantage in selling its products in international markets” [38]). There is some agreement among researchers that international competitiveness is strongly linked with productivity. According to the Organization for Economic Cooperation and Development (OECD) [38], when studying competitiveness, one should look at “the attributes and qualities of an economy that allow for a more efficient use of factors of production”. This is mirrored by Purwanto et al. [39], who said that competitiveness is “the extent to which firms in a particular region can compete with those elsewhere [and that] critical factors for competitiveness are those that determine the level of productivity in a region in relation to other regions”. These definitions fall in line with the World Economic Forum [40]: “competitiveness as the set of institutions, policies, and factors that determine the level of productivity of a country”. A similar definition of competitiveness was used by Schwab [41]: “attributes and qualities of an economy that allow for a more efficient use of factors of production.” One element that these definitions have in common is that they define international competitiveness as greater relative productivity and they do so by listing its determinants. Therefore, the logic that the more internationally competitive a country is, the higher the international comparative advantage (resulting from higher productivity) of its export will be is followed in this study. In other words, the more internationally competitive a country is, the higher its export relative to its import as compared to the same ratio for the world. This approach puts emphasis on the OECD’s [38] definition on international competitiveness provided earlier and the definition presented by Zhang [42]: “a country’s ability to compete internationally through expanding export capacity and upgrading export sophistication”. Presented logic also fits the conclusions of the Ricardo and the Heckscher–Ohlin theories (through lower opportunity cost and higher relative resource endowment to higher comparative advantage).

The relationship between energy security and international competitiveness is mainly investigated from the perspective of energy importing economies or mixed groups of countries. It is worth mentioning that the studies discussed here refer to international competitiveness on the macro level of a country and on the sole competitiveness of exporting industries or companies. Bilan et al. [43], analyzing Germany, on the one hand confirmed that energy security relates to international competitiveness, but on the other argued that even though Germany is highly competitive on international markets, it does not enjoy energy security due to the high share of natural gas import from Russia. Nawaz and Alvi [44], investigating Pakistan’s situation, claimed that energy security translates into the international competitiveness of domestic companies. Nyga-Łukaszewska and Chilimoniuk-Przezdziecka [7] checked how energy security affects competitiveness of export in a group of countries consisting of world’s biggest energy consumers (empirically testing export capacity function). The study confirmed the existence of a relationship between energy security and the ability of a country to compete with export of the capital goods on the international markets. Energy security influence over widely understood international competitiveness is not that often analyzed—unlike the isolated impact of energy prices (even called “energy competitiveness”) [45]. Under this notion, the authors recognized the influence of energy prices on export competitiveness [46,47] and treated energy prices (not energy commodities, like in Lieber [4], Klein [48] and McKinsey [49]) as a production factor (as suggested by Velthuijsen and Worrell [50]) determining international competitiveness. Scott [51] went beyond that by looking at price volatility and an adequate supply of fuels, all while assessing international competitiveness of low developed countries, suggesting that high energy costs and its inadequate supply influence their international competitiveness. A completely different approach to energy prices and international competitiveness is presented by Tvaronaviciene et al. [52]. The authors deliberately neglected energy prices in their analysis of competitiveness of specific industry branches exporting their goods to international markets (e.g., steel, wood). They did so by arguing that analysis of energy security should be narrowed down only to energy availability dimension. Tvaronaviciene et al. [52] concluded that in countries (without energy-intensive industrial branches), industry development is compatible with “energetically secure sustainable development and long-term competitiveness goals”. This study fits the existing research gap (see [7]) calling for broader empirical investigation of the energy security–international competitiveness link, especially in the group of countries divided by their GDP performance.

This study relies on the concept of developed countries. As Carbaugh [53] highlighted, advanced nations are primarily characterized by relatively high levels of GDP per capita, longer life expectancies and higher levels of adult literacy. It is worth noting that there is no coherent and universally accepted international policy standard on defining developed and developing nations. The OECD [54] highlighted that “there is no established convention for the designation of ‘developed’ and ‘developing’ countries or areas in the United Nations system”. That is why taxonomies of developed and developing countries offered by the OECD, World Bank or International Monetary Fund are not fully consistent, especially in the group of developing nations exporting crude oil. However, countries selected for this study are classified as advanced or high-income economies accordingly by the IMF [55], OECD [56] and World Bank [57].

3. Materials and Methods

The modeling approach employed in this study is based on Amoroso et al. [58], where the relative international competitiveness of Mexico—represented by the Revealed Competitive Advantage (RCA)—is modeled, including the Heckscher–Ohlin’s and Ricardo’s international trade hypotheses, which serve as a base for the model’s design. Amoroso et al. [58] relied on the methodology presented by Nunn [59] and Levchenko [60]. The latter study reflects the approach used in this study the most. Levchenko [60] investigated to what extent institutions affect trade and tried to operationalize an unmeasurable and elusive independent variable, such as institutional quality. The empirical aim of this study is to capture a similarly elusive concept of demand security as an explanatory component of international competitiveness. The approach presented by Amoroso et al. [58] is modified in two ways. First, the measure international competitiveness is done on a country level, not per industry. Second, the empirical part is extended to include a series of five models. All five models have a common part (explanatory variables representing the relative resource endowment and productivity differentials) and an individual part, which consists of one of the specific energy variables (energy security risk index, total natural resources rents, oil rents and coal rents) per model.

The path taken by in this study best fits the “competitive productivity” approach, which is listed by Ricardo et al. [61] as one of the “emerging approaches” to studying institutions and international competitiveness. The use of panel data (employed in this study) has been found by the same group of authors to be the most common methodology approach used, with generalized methods of moments (GMM) also being ranked high. It is worth noting that the use of GMM in this study was not possible due to a small number of cross-sections (3) relative to the number of periods (21, see [62] for details—three cross-sections leave us with no more than three instrumental variables). Given the design of the panel used in this study, the straightforward fixed-effects approach is the preferred option as with a large number of periods the “dynamic panel bias becomes insignificant, (…) the number of instruments in difference and system GMM tends to explode, (…) and [in case of a small number of cross-sections], the cluster-robust standard errors and the Arellano-Bond autocorrelation test may be unreliable” [63]. Reliance on factors of production (physical capital and labor force), its use and quality (e.g., human capital) as determinants of international competitiveness falls in line with recent works on this topic [64], which also note the continuing importance of the Ricardian and Heckscher–Ohlin theories in this aspect [65].

Since energy security as a concept is hard to capture with variables, the study uses a series of measures, which allowed to overcome the possible measurement bias. While operationalizing energy demand security, the argument of Sovacool and Mukherjee [26] is followed. They favor using reduced number of simple indicators in energy security studies over employing wide sets of data, convinced that it is more feasible to collect data for one of two simple indicators, as it may not be possible to collect data for a variety of metrics for particular countries. Energy variables included in this study aim at capturing energy security of exporters. Relying on Romanova’s [3], Bollino’s [9] and Dike’s [8] concepts of energy resource rents as a proxy of the price stability dimension in the case of exporters are included. This approach reflects the operational manner of Dike’s [8] economic dependence, which is calculated as value of energy export to GDP. This research also uses total natural resources rents as % of GDP, oil rents as % of GDP and coal rents as % of GDP. Such an approach has been used to test international competitiveness of other energy-exporting nations by, e.g., Quadah et al. [66] and Antoniadis [12]. “Rents” are defined as the difference between the value of resource production at regional prices and total costs of production. Unit rents are multiplied by the volume of energy production and referred to economy’s GDP [67]. Total natural resources rents are the sum of oil, natural gas, coal (hard and soft), mineral and forest rents [67]. Gas rents, on the other hand, are captured by the aggregated total natural resource rent variable. Additionally, a composite energy security risk index from the Global Energy Institute [68] is used to include a comprehensive measure. This index, as argued by Podbregar et al. [69], is mainly explained by the crude oil price and global coal reserves. Therefore, it mirrors the original understanding of energy security. Data used for energy variables stem from the Global Energy Institute database (index) [68] and the World Bank database (rents) [67].

There is no consensus on how to define international competitiveness [37]. Therefore, there is no unified way to measure it with some descriptive studies using multiple measures [70], while others use simple bilateral export to represent international competitiveness [64]. This research employs a ratio of studied country’s (i) export (X) to its import (M) divided by an alike ratio for the world (w) at time t (Equation (1)). Since this measure is based on the original RCA (e.g., Balassa [71]; Equation (2)), it is distinguished by adding *. This is not the first study to recognize the value along with the limitations of RCA, and like other authors, it is modified to better represent the modeled concept of international competitiveness (e.g., see [72,73]). For example, [65] developed what they call the “extended Balassa index”, which is a ratio of the traditional Balassa index calculated for exports to the same index calculated for imports. Importantly, [65] stated that this extended index “seems to be preferable from the theoretical point of view to the original version, given that the Ricardian and [Heckscher–Ohlin] theories state that not only export, but also import, depend upon competitive and [comparative] advantages”. Lastly, this measure of international competitiveness meets the three criteria for a measure of competitiveness set by Durand and Giorno [74]: it covers all sectors exposed to competition (i.e., it represents all goods traded that are subject to competition and only those goods) and it encompasses all markets open to competition and it is constructed from fully internationally comparable data:

$$RC{A}_{it}{}^{*}=\frac{\raisebox{1ex}{$X_{it}$}\!\left/ \!\raisebox{-1ex}{$M_{it}$}\right.}{\raisebox{1ex}{$X_{wt}$}\!\left/ \!\raisebox{-1ex}{$M_{wt}$}\right.}$$

(1)

$$RC{A}_{is}=\left(\frac{X_{is}/{{\displaystyle \sum }}_{s=1}^S{X}_{is}}{{{\displaystyle \sum }}_{i=1}^I{X}_{is}/{{\displaystyle \sum }}_{s=1}^S{{\displaystyle \sum }}_{i=1}^I{X}_{is}}\right)$$

(2)

To apply the concept of relativity, as present in the relative resource endowment theory and after Amoroso et al. [58], the final measure of relative international competitiveness is as in Equation (3):

$$\ln \left(\frac{\raisebox{1ex}{$X_{it}$}\!\left/ \!\raisebox{-1ex}{$M_{it}$}\right.}{\raisebox{1ex}{$X_{wt}$}\!\left/ \!\raisebox{-1ex}{$M_{wt}$}\right.}\right)-\ln \left(\frac{\raisebox{1ex}{$X_{Nt}$}\!\left/ \!\raisebox{-1ex}{$M_{Nt}$}\right.}{\raisebox{1ex}{$X_{wt}$}\!\left/ \!\raisebox{-1ex}{$M_{wt}$}\right.}\right)=\ln \left(RC{A}_{it}{}^{*}\right)-\ln \left(RC{A}_{Nt}{}^{*}\right)$$

(3)

where N represents Norway, which is the reference point compared against selected i economies. Countries included, next to Norway, in the research sample are: Canada, the Netherlands, the United States of America and Australia. High-income OECD countries being at the same time in the group of 10 leading world natural gas exporters have been selected for this study. Choosing the natural gas export as a proxy of energy export is a deliberate simplification as selected countries produce and sell on international markets other energy commodities as well. However, it is assumed that natural gas market is more regionally dispersed than crude oil and its importance as low-carbon fossil fuel will be rising in the future—especially when compared to other commodity traded internationally (hard coal). To keep the geographical diversity criterion, analyzed countries represent different continents. Norway and the Netherlands represent Europe, the USA and Canada represent North America and Australia represents Oceania. The Asian continent is not represented, as none of the countries from this region falls into a category of industrialized, world-leading gas exporter. This study does not include countries such as the United Kingdom (even though it is a European gas exporter), which in recent years lost its dominant market position, or Russia (being an Euroasian country), which is classified outside the developed economies group accordingly by OECD, IMF and World Bank.

Norway was selected as a reference point from the set of advanced economies being energy exporters (for which a full set of data was available) for a few reasons. Firstly, throughout the entire examined period (except for 2017), Norway has the lowest value of the energy security risk index mentioned earlier and the highest values on natural resource rents described above. Secondly, it is one of the most renowned energy exporters in the world, offering the most transparent regulatory policy, which as suggested by Kamprath [75] is “a role model for the 21st century energy policy”. Additionally, as third argument comes as a suggestion of the International Energy Agency [76] acknowledging that Norway’s oil and gas revenue management is commendable and may serve as a model for other countries to follow. That is confirmed by the fourth argument determining Norway’s choice, which is the position of the Norwegian sovereign wealth fund (SWF). The Government Pension Fund Global is, according to the Linaburg-Maduell Transparency Index [77], the most transparent SWF in the world, and as the Sovereign Wealth Fund Institute estimates, also the largest SWF in the world by the assets’ size [78]. Lastly, apart from being for many years a global benchmark for energy and fiscal policy, Norway leads in the energy transition ranking, setting the path for the low-carbon future [79], which from the perspective of a fossil-fuel exporter is a relatively unique strategy.

To include the relative resource endowment theory—and to abide by the above-mentioned notes on the link between international competitiveness and productivity—two factors of production are included in the model. The selection of these factors is supported by endogenous (e.g., Romer [80] and Aghion and Howitt [81]) and exogenous (e.g., Solow [82]) economic growth theories and fit the perspective on international competitiveness presented by Schwab [41], Purwanto et al. [39] and Zhan [42]. First, relative physical capital per worker endowment is multiplied by the fraction of physical capital to Gross Value Added (i.e., a proxy for physical capital intensity; Equation (4)):

$$\left[\ln \left(\frac{K_{it}}{L_{it}}\right)-\ln \left(\frac{K_{Nt}}{L_{Nt}}\right)\right]\frac{K_{it}}{GV{A}_{it}}$$

(4)

Second, relative endowment in human capital is multiplied by a ratio of a share of labor force with advanced education (a) to its average for the panel ($\overline{a}$, Equation (5)):

$$\left[\ln \left(H_{it}\right)-\ln \left(H_{Nt}\right)\right]\frac{a_{it}}{\overline{a}}$$

(5)

It is worth noting that in the original study by Amoroso et al. [58], relative endowment in human capital was multiplied by a ratio using wages instead of a share of labor force with advanced education, as is done here. This change was necessary due to the lack of comparable wage data for panel members.

Lastly, differences in productivity differentials are included according to Equation (6):

$$\ln \left(\frac{GV{A}_{it}}{L_{it}}\right)-\ln \left(\frac{GV{A}_{Nt}}{L_{Nt}}\right)$$

(6)

All energy variables (E) are included one-at-a-time in a series of five models and are represented by Equation (7):

$$\left[\ln \left(E_{it}\right)-\ln \left(E_{Nt}\right)\right]$$

(7)

where E is represented by a specific energy variable, one at a time: 1—Energy security risk index; 2—Total natural resources rents (% of GDP); 3—Oil rents (% of GDP); 4—Coal rents (% of GDP). A fifth model considering both energy security risk index and total natural resource rents at the same time was also run.

Descriptive statistics for constructed variables are provided in

Table 1. Descriptive statistics.

Statistic	$\mathbf{ln}\left(\mathit{R}\mathit{C}{\mathit{A}}_{\mathit{i}\mathit{t}}{}^{*}\right)-\mathbf{ln}\left(\mathit{R}\mathit{C}{\mathit{A}}_{\mathit{N}\mathit{t}}{}^{*}\right)$	$\left[\mathbf{ln}\left(\frac{{\mathit{K}}_{\mathit{i}\mathit{t}}}{{\mathit{L}}_{\mathit{i}\mathit{t}}}\right)-\mathbf{ln}\left(\frac{{\mathit{K}}_{\mathit{N}\mathit{t}}}{{\mathit{L}}_{\mathit{N}\mathit{t}}}\right)\right]\frac{{\mathit{K}}_{\mathit{i}\mathit{t}}}{\mathit{G}\mathit{V}{\mathit{A}}_{\mathit{i}\mathit{t}}}$	$\left[\mathbf{ln}\left({\mathit{H}}_{\mathit{i}\mathit{t}}\right)-\mathbf{ln}\left({\mathit{H}}_{\mathit{N}\mathit{t}}\right)\right]\frac{{\mathit{a}}_{\mathit{i}\mathit{t}}}{\overline{\mathit{a}}}$	$\mathbf{ln}\left(\frac{\mathit{G}\mathit{V}{\mathit{A}}_{\mathit{i}\mathit{t}}}{{\mathit{L}}_{\mathit{i}\mathit{t}}}\right)-\mathbf{ln}\left(\frac{\mathit{G}\mathit{V}{\mathit{A}}_{\mathit{N}\mathit{t}}}{{\mathit{L}}_{\mathit{N}\mathit{t}}}\right)$	$\left[\mathbf{ln}\left(\mathit{E}{1}_{\mathit{i}\mathit{t}}\right)-\mathbf{ln}\left(\mathit{E}{1}_{\mathit{N}\mathit{t}}\right)\right]$	$\left[\mathbf{ln}\left(\mathit{E}{2}_{\mathit{i}\mathit{t}}\right)-\mathbf{ln}\left(\mathit{E}{2}_{\mathit{N}\mathit{t}}\right)\right]$	$\left[\mathbf{ln}\left(\mathit{E}{3}_{\mathit{i}\mathit{t}}\right)-\mathbf{ln}\left(\mathit{E}{3}_{\mathit{N}\mathit{t}}\right)\right]$	$\left[\mathbf{ln}\left(\mathit{E}{4}_{\mathit{i}\mathit{t}}\right)-\mathbf{ln}\left(\mathit{E}{4}_{\mathit{N}\mathit{t}}\right)\right]$
i = Canada
Mean	−0.3656	−0.0002	0.0298	−0.5680	0.1301	−1.2431	−1.8084	3.4874
Maximum	−0.2286	−0.0001	0.0460	−0.3881	0.1973	−0.1182	−1.2657	8.2747
Minimum	−0.5165	−0.0002	0.0165	−0.7409	−0.0421	−2.7680	−4.4221	1.4378
Skewness	0.1792	0.0238	0.1695	0.3026	−1.4435	−0.8852	−3.3613	1.5402
Kurtosis	−0.6816	−1.4840	−1.6335	−1.3028	2.1199	3.7100	13.3641	1.9928
i = Netherlands
Mean	−0.3382	−0.0001	−0.0760	−0.4380	0.3368	−2.7016	−5.1141	5.5804
Maximum	−0.0606	0.0000	−0.0606	−0.1133	0.5461	−2.0319	−4.5514	10.9454
Minimum	−0.6189	−0.0001	−0.0886	−0.7302	0.1722	−3.9602	−5.7209	2.7753
Skewness	−0.1014	−0.8114	0.3642	0.1142	0.4471	−1.1541	−0.2439	1.1951
Kurtosis	−1.7227	0.2201	−1.4942	−0.1941	−1.7578	2.4380	−0.0412	1.4904
i = USA
Mean	−0.7624	−0.0001	0.0446	−0.2319	0.1703	−2.3108	−3.3666	4.1467
Maximum	−0.4137	−0.0001	0.0695	0.1239	0.2730	−1.1778	−2.7821	8.8856
Minimum	−1.1371	−0.0001	0.0243	−0.6046	−0.1184	−3.1235	−5.9367	2.2605
Skewness	−0.1252	−0.7393	0.3659	−0.0152	−1.7305	0.9614	−3.0186	1.5643
Kurtosis	−1.6413	0.2437	−1.3930	−1.6116	4.1789	3.6286	11.3275	2.3877

.

Data on export (in constant 2010 USD), import (in constant 2010 USD) and labor force (number of persons 15 years old and older supplying labor for production of goods and services), labor force with advanced (i.e., short-cycle, bachelor, master or doctoral or equivalent) education (% of total working-age population with advanced education), GVA (in current USD) and price levels used to adjust GVA (2010 = 100) were collected from the World Bank database [67]). Data on physical capital (in mil. 2011 USD) and human capital (index based on years of schooling and returns to education) were obtained from the Penn World Table, version 9.1 [83,84]. The one-year difference in base between physical capital and other variables is noted, but due to lack of large changes in price levels in such a short period, this is not expected to bias the obtained results. This study looks at the period of 1997–2017, which is dictated by data availability. Overall, there are 63 observations. Data availability has also limited the sample selection, excluding advanced energy-exporting economies such as Australia.

The general equation of the model used in this study can be written as Equation (8):

$$\begin{array}{cc}\ln \left(RC{A}_{it}{}^{*}\right)-& \ln \left(RC{A}_{Nt}{}^{*}\right)\hfill \\ & \hspace{1em} ={\beta }_0+{\beta }_1\left\{\left[\ln \left(\frac{K_{it}}{L_{it}}\right)-\ln \left(\frac{K_{Nt}}{L_{Nt}}\right)\right]\frac{K_{it}}{GV{A}_{it}}\right\}\hfill \\ & \hspace{1em} +{\beta }_2\left\{\left[\ln \left(H_{it}\right)-\ln \left(H_{Nt}\right)\right]\frac{a_{it}}{\overline{a}}\right\}+{\beta }_3\left[\ln \left(\frac{GV{A}_{it}}{L_{it}}\right)-\ln \left(\frac{GV{A}_{Nt}}{L_{Nt}}\right)\right]\hfill \\ & \hspace{1em} +{\beta }_4\left[\left[\ln \left(E_{it}\right)-\ln \left(E_{Nt}\right)\right]\right]+{ϵ}_{it}\hfill \end{array}$$

(8)

The models’ coefficients were estimated with Ordinary Least Squares with fixed effects selected based on the Hausman test’s results. Residuals were tested with the Wooldridge test for autocorrelation (Stata code: xtserial), modified Wald test for groupwise heteroskedasticity (xttest3), Pesaran’s test of cross-sectional independence (xtcsd, pesaran abs) combined with the Breusch-Pagan LM test (xttest2)–see [85], Levin–Lin–Chu (xtunitroot llc) and Harris–Tzavalis (xtunitroot ht) unit root tests and the Shapiro–Wilk (swilk) along with the Jarque–Bera (jb) test for normal distribution (see

Table 2. Model residuals’ tests summary.

Test for	Model 1		Model 2		Model 3		Model 4		Model 5
Code	xtgls, panel(hetero)	xtpcse	xtgls, panel(heter) corr(ar1)	xtpcse, corr(a1)	xtgls, panel(hetero)	xtpcse	xtreg, fe cluster()		xtgls, panel(hetero)	xtpcse
FE vs. RE	0.000 1		0.000 1		0.000 1		0.000 1		0.000 1
Autocorrelation	0.055		0.047		0.055		0.046		0.046
Heteroskedasticity	0.000		0.000		0.000		0.047		0.000
Cross-sectional indep. (Pesaran; BP-LM)	0.002	0.007	0.008	0.031	0.007	0.023	0.170	0.253	0.009
Normal distribution (JB; S-W)	0.010	0.000	0.008	0.000	0.012	0.000	0.007	0.000	0.008	0.000
Unit root (LLC; HT)	0.007	0.051	0.010	0.035	0.020	0.028	0.058	0.001	0.010	0.035
Prob > chi2	0.000	0.000	0.000	0.000	0.000	0.000	-	-	0.000	0.000
R-sq.		0.825		0.707		0.829	Within	0.798		0.26
							Between	0.581
							Overall	0.477

¹ “V_b-V_B is not positive definite”. This could not have been fixed with traditional measures, e.g., sigmamore.

).

Based on the results of the residuals’ tests and keeping in mind observations of Hoechle [86], variations of Generalized Least Squares (GLS) and Panel Corrected Standard Errors (PCSE) were used. Since GLS (xtgls) tends to produce overoptimistic standard errors, an inclusion of PCSE (xtpcse) should be seen as a results’ robustness check. In Model 4 (due to no cross-sectional dependence), robust standard errors were obtained with the cluster() option in Stata.

Lastly, the Kao test for cointegration has been carried out (xtcointtest kao) for all five models (see

Table 3. Results of the Kao test for cointegration.

Test	Model 1	Model 2	Model 3	Model 4	Model 5
Modified Dickey–Fuller t	0.008	0.143	0.223	0.001	0.051
Dickey–Fuller t	0.018	0.108	0.101	0.006	0.074
Augmented Dickey–Fuller t	0.007	0.002	0.026	0.037	0.001
Unadjusted modified Dickey–Fuller t	0.050	0.066	0.041	0.003	0.065
Unadjusted Dickey–Fuller t	0.028	0.077	0.041	0.008	0.080

). Generally, the null hypothesis of no cointegration can be rejected for all individual statistics for all models at a 10% level of statistical significance. For models 2 and 3, the Modified Dickey–Fuller t provides the opposite and the Dickey–Fuller t borderline results. Given that the used variables are constructed using many processes and that most results support rejection of the null hypothesis of no cointegration, it is possible to say that there exists a long-run equilibrium of the studied process. The obtained Kao test for cointegration results may be biased due to the test demand for large-N-large-T-panel data, which this dataset does not fulfill.

Prior to moving further, it is important to recognize the limitation of the employed approach. Namely, since variables in the model are a result of transformations (e.g., relative physical capital per worker is a combination of five variables) it is impractical to interpret estimated coefficient in the traditional per-unit change fashion. Instead, their interpretation is limited to their statistical significance and the direction of change (i.e., the sign).

4. Results

Coefficients of energy-independent variables are consistently not statistically significant within and across all models (

Table 4. Models’ results.

Coefficient	Model 1		Model 2		Model 3		Model 4	Model 5
Code	xtgls, panel(hetero)	Xtpcse	xtgls, panel(hetero) corr(ar1)	xtpcse, corr(a1)	xtgls, panel(hetero)	xtpcse	xtreg, fe cluster()	xtgls, panel(hetero)	xtpcse
Physical capital per worker	−2767.6 ***	−3116.7 ***	−1651.13 ***	−1820 ***	−2587 ***	−2910 ***	−2584.75 **	−2823.8 ***	−3244.82 ***
Human capital	−3.76 ***	−4.16 ***	−3.57 ***	−3.72 ***	−4.38 ***	−4.58 ***	−12.26 **	−3.73 ***	−4.11 ***
Labor productivity	−0.51 ***	−0.45 ***	−0.31 ***	−0.30 ***	−0.43 ***	−0.37 ***	−0.25 **	−0.52 ***	−0.47 ***
Energy security risk index	0.03	−0.08						0.02	−0.08
Total natural resources rents (% of GDP)			0.04 **	0.03				−0.01	−0.01
Oil rents (% of GDP)					0.03	0.03
Coal rents (% of GDP)							0.02
const	−1.05 ***	−1.05 ***	−0.73 ***	−0.76 ***	−0.90 ***	−0.9 ***	−1.02 ***	−1.08 ***	−1.10 ***

*** p-value < 0.01, ** p-value < 0.05.

). Although in Model 2 xtgls specification, total natural resources rents have a p-value of 0.039, in the light of results of other models, this can be assigned to the overoptimistic standard errors expected of this specification–as mentioned earlier. This means that international competitiveness of studied economies is not determined by the energy security index or any of the studied energy rents. From the methodological perspective, the obtained results show that whether a comprehensive or a fuel-specific measure of energy security is used, the results do not change.

Consistently (within and across models), coefficients of relative resources endowment and labor productivity differentials are statistically significant and negative, which speaks to the consistency of obtained results. Negative coefficients mean that when relative endowment in physical capital of Norway increases—$\left(\frac{K_{Nt}}{L_{Nt}}\right)\uparrow \Rightarrow \left[\ln \left(\frac{K_{it}}{L_{it}}\right)-\ln \left(\frac{K_{Nt}}{L_{Nt}}\right)\right]\downarrow$—relative to countries’ i, the international competitiveness of Norway will decrease—$\left[\ln \left(RC{A}_{it}{}^{*}\right)-\ln \left(RC{A}_{Nt}{}^{*}\right)\right]\uparrow$. A parallel interpretation applies to the relative endowment in human capital and labor productivity. Similarly, $\left(\frac{K_{it}}{L_{it}}\right)\uparrow \Rightarrow \left[\ln \left(\frac{K_{it}}{L_{it}}\right)-\ln \left(\frac{K_{Nt}}{L_{Nt}}\right)\right]\uparrow$ leads to a fall in relative international competitiveness of economies compared to Norway, $\left[\ln \left(RC{A}_{it}{}^{*}\right)-\ln \left(RC{A}_{Nt}{}^{*}\right)\right]\downarrow$. Since all analyzed countries are world-leading energy-exporting developed economies, a similar interpretation of the estimated coefficients also speaks to the results’ consistency. In this study, Norway is chosen as a background for the analysis of the demand security influence over international competitiveness of high-income energy-exporting OECD countries. Choosing Norway as a benchmark is justified by its long-standing position as an exporter of energy resources. High institutional quality (e.g., rule of law) in energy markets also makes it a good example for other countries. The relevance of the obtained results extends to other analyzed economies as well.

The empirical procedure shows that the relative international competitiveness of Norway is determined by its relative physical and human capital endowment and its relative labor productivity. Energy security explanatory variables are not statistically significant. Therefore, the research hypothesis claiming that relative energy demand security has a positive and a statistically significant impact on the relative international competitiveness of developed, energy-exporting economies has been rejected. Furthermore, the Heckscher–Ohlin’s and Ricardo’s international trade theory have been shown to be applicable when modeling international competitiveness of studied economies. While accounting for cross-sectional dependence with both GLS and PCSE, there were no significant differences in terms of statistical significance of the coefficients between the two methods.

5. Discussion

This study focuses on the role of energy security as a determinant of the international competitiveness of developed, energy-exporting economies. This carefully selected group of countries, as shown by the above model, has not built its international competitiveness on energy demand security. In the case of analyzed countries, resource rents or composite energy security risk index do not play a role in their capacity to compete on international markets. This conclusion contradicts Novikau’s [15] argument of the sole importance of energy demand security for exporters and energy supply security for importers. Adding to this, the results of previous research [7], in which authors found a positive relationship between energy supply security and export’s competitiveness (among others: the USA and Norway), it may seem that for advanced energy exporters the idea of “energy competitiveness”—introduced by Zachmann and Cipollone [45]—is relevant. That, again, proves the high “contextuality” of general energy security research [87]. According to this idea, introduced by Winzer [87], energy security is a country-specific phenomenon. To some extent, this study might fit into the general debate on the Dutch disease phenomenon (sometimes also called the resource curse) as high dependence and reliance on energy export is its trigger. This concept, first described by Corden and Neary [88], is applicable to all resource-rich economies suffering from an incorrect allocation of production factors within an economy’s sectors. Various countries across the world have been suspected to suffer from the Dutch disease symptoms, e.g., Chile (due to copper exports) [89], Russia (due to oil and gas exports) [90] and a developed economy such as Australia (due to coal and gas exports) [91]. As a results of becoming an energy exporter, the US has also been recently investigated for Dutch disease symptoms. However, as Allcott and Keniston [92] prove, oil and gas booms in the US did not crowd out the manufacturing sector, which is one of the prime disease symptoms. On the contrary, the sector exhibits an overall growth. Canada, another resource-endowed country from the sample used in this study, since 2002, has suffered from the Dutch disease symptoms such as appreciation of the Canadian dollar and a decline in size of its manufacturing sector. However, whether Canada contracted the Dutch disease or not remains a controversial issue. One of the studies offering a possible explanation is Beine et al. [93], who concluded that Canada, due to its heavy reliance on the US as a trading partner, suffers rather from a “Canadian disease” or a “Dutch Affair” [94] than the Dutch disease. The US–Canadian trade “gravity” causes the Canadian dollar to appreciate whenever the US dollar depreciates, making the Canadian manufacturing sector less competitive on the international markets. Coulombe [94] added that the “Dutch Affair becomes a disease in the long run when the non-renewable resource is depleted, and the manufacturing base is gone” in Canada. However, that is yet to be observed. Interestingly, Norway is one of those resource-rich countries that warded off the Dutch disease symptoms. According to Ramirez-Cendrero and Wirth [95], successful and well-developed oil and gas policy prevented Norway from contracting the resource curse. This idea supports previous studies of Mehlum et al. [96], which contrasts the work by Sachs and Warner [97], by concluding that the quality of institutions is dependent on whether resource-abundant country will or will not experience a resource curse. Mehlum et al. [98] stated that “grabber friendly institutions and resource abundance produces a growth trap. Producer friendly institutions, however, help countries to take full advantage of their natural resource abundance”.

Based on the above discussion and obtained results, the same quality of institutions might be a possible explanation for insignificance of resource-rent variables in all econometric models. One may expect that in case of resource-rich developed countries, it is the fiscal policy and regulatory environment that firstly catalyzes energy-industry innovations and secondly introduces frameworks separating resource-prices and their budgetary dependency. Comparison of the position of Russia and Norway in the Global Competitiveness Index rankings proves that R&D spending supports the Norwegian industry innovations.

Econometric insignificance of resource-rent variables might be exacerbated by the economic conditions of the countries in the sample. Because all analyzed nations are relatively well diversified and independent of natural resource incomes, the significance of oil/coal/gas rents is diminished. Industrialized nations have already achieved economic diversification, which is suggested by Romanova [3] for energy exporters. Therefore, their economic performance is not vulnerable to any volatility in energy prices. Additionally, well-designed institutional frameworks protect these countries from transmission of price increases and decreases onto their economies’ performances. This explanation is also applicable to the irrelevance of the energy security risk index, especially when it is mainly influenced by prices of resources and the availability component.

Another possible explanation of this research results is a combination of effects connected with the composition of the countries’ sample and methodological framework. Amoroso et al. [58] estimated the relative differences in the respective variables between countries using one of the countries as a baseline. Even though it is assumed that a homogenous sample will be beneficial for study results, it turns out that great economic similarity might have negatively affected any possible differences in energy security approaches. Comparing developed Norway with high energy policy standards along with the US, Canada and the Netherlands (which enjoy similar conditions) did not reveal any characteristics that might have been captured statistically. Thus, any effort to broaden the developed or even developing countries’ sample without data limitations might be an interesting next research step.

By proving the lack of statistical significance of energy rents in shaping international competitiveness of energy exporters, this study fits into the general debate on the sources of competitiveness. It has been observed that in the group of the OECD countries, and EU-27 in particular, price competitiveness does not affect their outcome competitiveness [98]. That is why “narrow focus on the price component of competitiveness neglects other aspects of the concept that are likely to be particularly important for high-income economies” [96]. A possible source of international competitiveness of developed energy exporters might be energy-related innovations. As the OECD study [99] suggested, innovations in oil and gas sectors of countries such as the US, Canada and Norway are mainly driven by environmental protection concerns, not price considerations. Additionally, these innovations are carried out largely by an industry with limited government involvement. Therefore, any effort to capture the possibility of hydrocarbons’ sector innovation influence on international competitiveness would have to be carried out on a company level. Norwegian Equinor is a suitable example for such an analysis, which is depicted as a possible next research step.

6. Conclusions

This study reflects on two fiercely discussed topics: international competitiveness and exporter’s energy security on the country level. The main objective of this study was to examine energy demand security and its influence on international competitiveness. This study sheds a light on a novel approach to energy security while taking into account the perspective of energy exporters. The advantage of this research is that, in contrast to other studies, empirical analysis is embedded within the international trade theory framework. This is a novel approach in energy security studies.

Research focuses on the scientifically underreported phenomenon of energy security of highly developed energy exporters. Thus, Norway is put at the center of this analysis, along with the US, Canada and the Netherlands. This research proves that energy security does not play a role in the shaping of international competitiveness of highly developed energy-exporting countries. Sources of their international competitiveness are linked to other than resource rents and energy availability factors. Rejecting the research hypothesis is followed by a set of theoretical and empirical implications.

Regarding the specific results of the econometric model, the null hypothesis of no impact of relative physical capital per worker, relative human capital and relative labor productivity on relative international competitiveness have been rejected at a 5% level of statistical significance. Interestingly, the obtained negative sign of these coefficients leads to the conclusion that relative international competitiveness between developed economies, which are also net energy exporters, is not positively determined by the relative abundance of traditional factors of production.

From the theoretical point of view, this study highlights research needs on energy security of exporting nations. Even though in the last few years there has been a growing trend in that research field, the concept still lacks a detailed description supported by an in-depth econometric analysis. The majority of contemporary studies concerning energy security looks at the phenomenon from the point of view of the importing nation, whereas the importance of demand security has been to some extent neglected. One possible explanation behind it might be the historical development of energy markets, in which energy security (of importing nations) emerged as a problem in the late 1970s and early 1980s. Being a relatively new phenomenon, this concept required scientific attention that was not shared with demand security. With new infrastructural investments being commissioned across the world (new pipelines, e.g., Nord Stream I and II) and new players entering energy export markets (e.g., US), questions of stability of incomes for exporters and their investment rights have emerged. This refers more often not only to developing or transition economies but also developed ones.

That is why energy demand security of highly developed energy-exporting countries poses a significant empirical challenge. Firstly, the number of such countries is relatively low around the world. Therefore, any quantitative research (especially a panel study) must be carefully designed keeping in mind data limitation problems. Secondly, in this group of countries, other determinants might be relevant in the energy markets analysis, such as the quality of institutions or technological progress. Inclusion of such soft variables on the macroeconomic level proves to be a challenge. Additionally, from the operational point, including all possible variables, would run the risk of model’s overspecification and severe multicollinearity. One may expect that demand security in the group of highly developed nations can be associated with R&D spending on energy-related technologies or informal institutions protecting countries from the “Dutch disease”. It is possible that in this group of nations, any empirical investigation should rather focus on variables describing the export of advanced energy-related technologies with high-tech content, instead of looking at resource rents or even the quantity of energy exported. That is why investigating the influence of the above-mentioned topics on international competitiveness of developed countries is recommended as one of the next research steps.

Study findings are also of importance for policymakers, as they bring two possible policy deliverables. First, if (as suggested by our results) energy demand security does not play a role in shaping of international competitiveness of high-income countries, decisionmakers in those nations may consider focusing solely on energy supply security. This idea to some extent contradicts Novikau’s approach [36] stating that only demand security matters for energy exporters. It may suggest that energy policy agenda can be diversified with other than demand security challenges and possible ways of embracing them. That is why when developing energy strategies, exporters should focus on the most pressing policy areas. Allocating resources in these policy areas increases the chances for success. Possibly, the most urgent problem nowadays is socially and economically fair energy transition. Moving towards a low-carbon economy structure poses additional challenges related to maintaining international competitiveness as well as building new comparative advantages in global supply chains. Addressing those problems today, by prioritizing and supporting “green” innovations, may help advanced economies stay competitive in the future. Second, advanced economies exporting energy resources, being already insulated from energy price volatility through sovereign wealth funds, are not bothered with the way energy rents are managed. Since these economies do not rely on energy rents in shaping their international competitiveness, their sources of competitiveness are located in other than energy-exporting industries. That is why policies aimed at increasing international competitiveness should rather focus on sectors other than energy-export sectors as a mean of the most efficient resource allocation.

Like any study, this study has its limitations, which might be a source for further research in this field. First, a narrow spectrum of energy-exporting, world-leading developed economies is tested. A future study might be extended to include either a larger panel of developed economies or to examine the role of energy export in determining relative international competitiveness of Norway (or another similar economy) against developing economies. The aim of this study was not to examine all sources of international competitiveness of Norway, but to focus on the role of energy export in this area. Nevertheless, the econometric model we used has shown that increases in endowment of traditional factors of production do not positively impact the studied dependent variable within the group of developed economies. Therefore, further studies could search for factors that impact the relative international competitiveness of Norway or other high-income energy exporters. Second, this study employs the demand security concept in its economic dimension, and the model tries to operationalize it. Even though reliable data sources are used, offering full time and sample coverage, stability of export incomes might be operationalized in other ways and, thus, bring different results. That is why deepening scientific investigation on demand security is suggested as a possible next research direction.