Influence of Oil Factor on Economic Growth in Oil-exporting Countries

Traditional types of energy resources, among which oil is primary, have a huge influence on economic development in various manifestations. In our research, we discuss dependence of economic growth in oil-exporting countries on first-order factors, such as oil prices, oil production, and structural shift in a share of oil exports. To review 2005-2019, we chose leading oil-exporting nations, including the OPEC members (Iraq, Iran, Libya, Saudi Arabia, and Nigeria), as well as countries outside this group (Russia, Kazakhstan, Azerbaijan, and Norway). Having used statistical and regression techniques, we confirm that the correlation between oil price fluctuations and economic growth raises with the scale effect. Larger economies (by absolute GDP in oil-exporting nations and hydrocarbon production) might generate more intensive economic growth from positive changes in oil prices than small economies. Also, the OPEC members show the strongest structural shifts in the change of the share of oil exports than the other countries. Resulting regression dependences led us to the conclusion that in oil-exporting countries, economic growth largely depends on an increase in oil prices and changes in oil production rates. But contrary to the expectations saying that in time of downturns, as far as global oil prices fall, nations will reduce oil exports, we observe the reverse picture – changed role of the price factor. We explain this with the keeping income-balance strategy, where the oil price parameter is a tool to choose ideal production. Our findings show that with the 1%-increase in oil production, the countries under considerations (the OPEC members) might reach GDP from 0.0367% (Iraq) to 0.437% (Libya). For countries outside OPEC, such a GDP growth might be more intensive. For instance, in Russia, it might be up to 1.559%. This also points out that the joint coordinating policy in oil production affects the economic growth potential.


INTRODUCTION
Efficient functioning of national economies largely depends on their structures, proportions of leading sectors, and performance. Economy performance evaluation is of particular relevance for the nations that might be classified as non-diversified in terms of a market power of one or a number of related sectors (Kuboniwa, 2012;Figueroa and Calfucura, 2010;Litau, 2018;Kheyfets and Chernova, 2021). This statement is also true for the nations that specialize in exports of natural energy resources, such as oil and gas (Xiong et al., 2015;Al-Rawashdeh et al., 2013;Ji et al., 2014;Illig and Schindler, 2017).
The upward trend in the cost of energy resources has dramatically changed a nature of international relations. They did not only become a commodity, but also a tool for influence in geoeconomy and geopolicy. The balance of power has shifted towards the developing countries that export large volumes of oil and gas. Despite steady growth in alternative (renewable) energy sources, they are not yet able to meet needs of global economy in full (Kuzmin et al., 2019). However, the global oil market has significant fluctuations. So, for 9 months of 2020, the fall in prices of Brent crude oil was 67.9% (Fusion Media Limited, 2021), whereas a year earlier, growth in oil prices was over 22. 2% (Fusion Media Limited, 2021). Such variability, in turn, causes significant economic shocks in the countries that specialize in oil production and exports as dependence on market conditions creates many risks for economic growth in such countries. Therefore, it remains important to clarify a nature of the influence of the dominating oil export on economic development.
These circumstances governed the purpose of our research, which we state as evaluating an influence of oil prices, oil production, and structural changes in exports on economic growth of oil-exporting countries. To do this, we solved the following tasks: explore export orientation in oil-exporting countries (define a share of oil exports of total exports in the countries under consideration), analyse structural changes in oil exports, evaluate the interrelations between oil prices, changes to oil exports share, oil production, and economic growth rates (GDP).

LITERATURE REVIEW
In the 1980-ies-1990-ies, people mostly believed that natural resource endowment slows down economic growth (Gylfason et al., 1999;Sachs and Warner, 1995). Auty (2001) shows that such large resource-based economies, as Russia, Nigeria or Venezuela have lower rates of GDP growth than many countries that possess significant natural resources (Japan, South Korea, Ireland, etc.). The concept of resource curse or resource surplus paradox has become widely known. The natural resource surplus might cause very diverse effects: from low budget deficits and high living standards of the population to weak institutions and slowdown in economic growth rates (Sim, 2013).
Regarding economies in transition, Kronenberg (2004) points out to the essential inverse correlation between availability of raw materials and economic growth. Esanov et al. (2001) made similar conclusions for some CIS countries. Turkmenistan and Uzbekistan are illustrative cases. Both countries, as well as Azerbaijan and Kazakhstan, are described with authoritarian rule and monopolization in economic environment. The resource rent provides for functioning of a kind of political and economic oligopoly that impedes economic reforms and growth.
The need to maintain export earnings of countries means the influence of the price factor on export positions. The influence of the price factor on countries depends on aggregate structures of exports and imports. For instance, many developing countries export raw materials, but import energy resources (Spatafora and Tytell, 2009). Accordingly, the price boom in commodity markets does not necessarily lead to significant countries improvement. At first glance, declined commodity prices might neutralize the trend towards excessive strengthening of national currencies in exporting countries and thus neutralize the Dutch disease threat. However, functional dependencies are thought to be much more complicated. Ftiti et al. (2016) proposed the econometric methodology from the evolutionary spectral analysis to distinguish between short-term coincidence and long-term dynamics. Their findings show that oil price fluctuations have both medium and short-term influence on economic indicators in the OPEC countries. At the same time, medium-term effects were significantly larger than short-term ones. Umar et al. (2017) identified cause-and-effect relations between oil price volatility and economic performance in Nigeria. Based on vector autoregressive patterns, they found the close connection between oil price fluctuations and GDP growth. Gross domestic product was considered as a dependent variable with such main explanatory factors, as oil price, direct foreign investments, total export, parameters of trade openness and level of human capital. However, in time of simulation efforts, Umar et al. found no correlation between variables in the long term horizon of their review. They concluded that to keep sustainable economic growth, we need to keep less volatile world oil prices as this parameter is indeed a significant factor in terms of influence on economic growth in oil-exporting countries. Shibinskaya and Burakova (2017) explored dependence between oil prices and growth in U.S. economy in the long term. For an analysis, they applied the VAR method, which shows that abrupt changes in oil prices in part suppressed GDP growth.
Since Russia is one of the leading oil and gas exporters, numerous research describe dependence of Russia's GDP on levels of oil prices. So, for their calculations, Shestukhina and Fomin (2016) applied the least squares method. They found the close relationship between the indicators under consideration. Rau (2017) also shown the effect of the changed oil prices on GDP.
The finalised literature review shows ambiguity in views of researchers on a nature of the influence of national resource bases on economic indicators of development. There are still in place unsolved questions that relate in particular to transformation of macroeconomic policies of raw-exporting countries in terms of abrupt price changes, observable in the active phase of the COVID-19 pandemic. Thus, keeping economic development sustainable is an adaptational task aimed at structural changes. In this context, we try to clarify the strength of the influence (sensitivity) that oil prices make on economic growth in some countries, considering the latest trends in the market environment.

MATERIALS AND METHODS
Economic growth models are based on various approaches, from production functions and vector autoregression to non-parametric design (Jiménez-Rodríguez and Sánchez, 2005). However, in our research, we apply the aggregate indicator of gross domestic product as a simple alternative to estimated dynamics of economic growth. Coming from the finalised literature review and own observations, we believe that the relationship between fluctuations in oil prices and economic growth is stronger in large economies (by absolute GDP of oil-exporting nations and hydrocarbons production).
To make the review, we chose leading oil-exporting nations. Among them, there are the OPEC countries (Iraq, Iran, Libya, Saudi Arabia, and Nigeria), as well as the countries outside this group (Russia, Kazakhstan, Azerbaijan, and Norway). We applied the heterogeneous sample of countries (that includes non-OPEC countries) to reduce factors of mutual influence within the OPEC group. The sample was also promising as the countries varied by level of socio-economic development.
The research of oil price and production influence on economic growth included three stages: 1. At the first stage, we explored foreign trade focus in the countries under consideration, for which we applied the export quota (EQ) indicator: 2. At the second stage, we calculated the share of oil exports of total exports in the countries, as well as analysed changes to structural shifts in this share. To process the data, we applied the structural-dynamic analysis. It assumes that we find individual and summarising indicators of structural changes (using linear and root-mean-square coefficients of absolute and relative structural shifts with variable or fixed comparison base). We calculated the structural indicator using the formula by Sivelkin and Kuznetsova (2002). 3. At the third stage, we reviewed the relationship between the share of oil exports in total exports and rates of economic growth (GDP). To carry out the correlation analysis, we applied the multiple linear regression. It assume the calculation of the linear dependence in the group of independent variables.
To estimate the relationship between structural shifts in the share of oil exports, changes in oil prices, and GDP dynamics rates in oil-exporting countries, we found the correlation coefficient (Ishkhanian and Karpenko, 2016): Where n is size of the population under consideration (sample size); x , y are average values of parameters; σ x 2 , σ y 2 are dispersions of time-series parameters; and σ x , σ y are root-meansquare (standard) deviations of parameters.
The multivariate regression model might look like the following: where x 1 is change in structural shifts in the oil exports share; x 2 is oil production growth rates; x 3 are oil prices in a respective country.
We tested overall quality of the multiple regression equation using Fisher's criterion. H 0 is the hypothesis for statistical insignificance of the regression equation and strength of the relationship index. If F calc ≥F table , we reject H 0 hypothesis and recognize significance and reliability of estimated parameters of the regression equation.
The research information base included data on socio-economic situations of countries in the sample, i.e. GDP in raw numbers, GDP per capita, data on proven oil reserves and production, oil exports in kind (barrels), value of total exports and value of exports of the oil they produce, as well as the data on oil national demand (consumption).
Research assumptions and constraints included cost characteristics of oil. Allowing for the cost of oil, we applied oil grades prevailing in exports.

RESULTS
The research background includes the review of socio-economic conditions in the countries under consideration. The sample includes small-size countries with the population of not more than 10 million people (Norway, Azerbaijan, and Libya) and large-size countries with the population of more than 50 million people (Iran, Russia, and Nigeria). See the indicators that describe levels of socio-economic development of these countries in Table 1.
In 2019, in most of the countries under consideration, the export quota was over 0.3 (Iraq, Libya, Kazakhstan, Azerbaijan, Norway, and Saudi Arabia). Export was not so significant for Iran, Nigeria, and Russia. Many countries under consideration were highly dependent on oil exports as its share in total value of exports was over 70% (Iraq, Libya, Saudi Arabia, Nigeria, and Azerbaijan). In Iran, the share of oil exports of total export was about 28%, while in Russia, it was 29%, in Kazakhstan, it was 68%, and in Norway, it was 39%. See the data on dynamics of oil production in 2005-2019 in Table 2.
In 2005/2019, Iraq (+246.9%), Kazakhstan (+57.1%) and Azerbaijan (+54.8%) had the highest oil production growth rates. Some countries had a decrease in oil production: Libya (−35.2%), Nigeria (−26.6%), and Norway (−44.8%). At the same time, by comparing the data on dynamics of changes in production and oil exports in kind, we might observe that oil exports in Iran increased 2.69 times, and in Azerbaijan -2.57 times. In Russia, the increase in oil exports was only 8.6%. In the rest of the countries under consideration, the growth in oil exports declined. This might be due to increased national consumption. See graphical representation of average annual growth rate (crude oil production and exports) in Figure 1.
Despite the slight decrease in oil exports in kind, in value terms, almost all countries had growth (    Based on available data, let us calculate the structural indicator of the share of crude oil exports in total exports (Table 4) and examine its change in dynamics. For this, we calculated the individual index of absolute structural shifts with constant ∆ 1 (Table 4) and variable comparison base ∆ 2.
In 2005-2019, we found the largest structural changes in exports in such countries, as Iran (decrease by 60.94% in the share of oil exports), Nigeria (decrease by 29.63% in the share of oil exports), and Libya (decrease by 12.27% in the share of oil exports). In Azerbaijan, growth in the share of oil exports was 32.82%, in Kazakhstan +13.14%, and in Norway + 10.29%. Russia slightly reduced the share of oil exports (by 5.64%).
In Thus, the OPEC member states show the strongest structural shifts in the changed share of oil exports than other oil-exporting countries.
In the review of time periods of change structural shifts, we might mention that in 2008, three out of nine countries under considered shown the increase in the share of oil exports (Iran, Libya, and Azerbaijan). This was accompanied by significant growth in prices on the oil market. This what the reason for changes in oil exports.
The opposite fluctuation was in 2009, when oil prices were strongly corrected. Iran Light had the price reduction of 35.29%, Brega had the price reduction of 36.65%, Arab Heavy' price decreased by 31.63%. As a result, substantial reduction in the share of oil exports was showed by Iran, Libya, and Saudi Arabia (countries most sensitive to the price factor).
The influence of the price factor gradually escalated. In 2015, five out of nine countries under consideration reduced the share of oil exports (Iran, Saudi Arabia, Russia, Kazakhstan, and Azerbaijan). At that time, the decrease in the price of Iran Light, Arab Heavy, Urals, and Brent was 47-49%.
Behaviours of exporting countries in the oil market have changed in recent years. Maintaining the balance of incomes has become a  priority (when the price factor is compensated by supply volumes). Therefore, in 2019, five out of nine countries increased the share of oil exports (Iraq, Libya, Kazakhstan, Azerbaijan, and Norway), even though in that time oil prices were in decrease. Thus, the cost of Brent, Brega, Urals, and Ekosfisk reduced by 8-9%. This was influenced by double term extension of the OPEC+ agreement, trade war between the USA and China, peak growth in shaly production in the United States and other countries, tankers seizures, attacks on oil treatment plants in Saudi Arabia, and sanctions against Iran and Venezuela (Deynego, 2020).
The analysis performed makes it possible to move on to the next research stage. The raw data (to make regression model) included GDP growth rates of 2005-2019 in the countries under consideration (Table 5).
Based on the earlier estimated structural shifts in the share of oil exports, oil production growth rates, as well as dynamics of oil prices, we made regression dependencies for each country.
The regression equation for Iraq is as follows: Y=−0.4132+0.207 4x 1 +0.4267x 2 +0.04098x 3 , We tested statistical significance of the equation using the determination coefficient and Fisher's criterion.
Overall variability of GDP in Iraq is due to 55.1%-change to x i factors by R 2 . Tabular value Fkp(3,10)=3.71, Ffact=4.09. When Ffact>Fkp, the determination coefficient is statistically significant and regression equation is statistically reliable. From the highest coefficient β we conclude that the greatest influence on GDP in Iraq is made by the magnification factor of oil production growth rates.
The regression equation for Iran is as follows: Y=0.8186+0.0785 7x 1 +0.2096x 2 +0.06918x 3 . We tested statistical significance of the equation using the determination coefficient and Fisher's criterion.
Overall variability of GDP in Iran is due to 18.7%-change to x i factors by R 2 . Tabular value Fkp(3,10)=3.71, Ffact=0.77. When Ffact<Fkp, the determination coefficient is not statistically significant and regression equation is not statistically reliable. From the highest coefficient β we conclude that the greatest influence on GDP in Iran is made by the magnification factor of oil production growth rates.
The regression equation for Libya is as follows: Y=−0.06079+0.4 78x 1 +0.4373x 2 −0.2056x 3 . We tested statistical significance of the equation using the determination coefficient and Fisher's criterion.
Overall variability of GDP in Libya is due to 92.7%-change to x i factors by R 2 . Tabular value Fkp(3,10)=3.71, Ffact=42.29. When Ffact>Fkp, the determination coefficient is statistically significant and regression equation is statistically reliable. From the highest coefficient β we conclude that the greatest influence on GDP in Libya is made by the magnification factor of oil production growth rates.
The regression equation for Saudi Arabia is as follows: Y=0.0816 +0.1037x 1 +0.3364x 2 +0.00189x 3 . We tested statistical significance of the equation using the determination coefficient and Fisher's criterion. Overall variability of GDP in Saudi Arabia is due to 70.8%-change to x i factors by R 2 . Tabular value Fkp(3,10)=3.71, Ffact=8.09. When Ffact>Fkp, the determination coefficient is statistically significant and regression equation is statistically reliable. From the highest coefficient β we conclude that the greatest influence on GDP is made by the magnification factor of oil production growth rates.
The regression equation for Nigeria is as follows: Y=1.96 62+0.10563x 1 +0.033667x 2 +0.01268x 3 . We tested statistical significance of the equation using the determination coefficient and Fisher 's criterion. Overall variability of GDP in Nigeria is due to 7.0%-change to x i factors by R 2 . Tabular value Fkp(3,10)=3.71, Ffact=0.25. When Ffact<Fkp, the determination coefficient is statistically non-significant and regression equation is not statistically reliable. From the highest coefficient β we conclude that the greatest influence on GDP is made by the magnification factor of the individual structural shift in the oil export share.
The regression equation for Russia is as follows: Y=−0.8788−1.2 874x 1 +1.5587x 2 +0.1846x 3 , We tested statistical significance of the equation using the determination coefficient and Fisher's criterion.
Overall variability of GDP in Russia is due to 68.6%-change to x i factors by R 2 . Tabular value Fkp(3,10)=3.71, Ffact=7.28. When Ffact>Fkp, the determination coefficient is statistically significant and regression equation is statistically reliable. From the highest coefficient β we conclude that the greatest influence on GDP in Russia is made by the magnification factor of oil prices growth rates.
The regression equation for Kazakhstan is as follows: Y=2.7 881+0.04114x 1 −0.0148424x 2 +0.07488x 3 , We tested statistical significance of the equation using the determination coefficient and Fisher's criterion. Overall variability of GDP in Kazakhstan is due to 38.7%-change to x i by R 2 . Tabular value Fkp(3,10)=3.71, Ffact=2.11. When Ffact<Fkp, the determination coefficient is statistically non-significant and regression equation is not statistically reliable. From the highest coefficient β we conclude that the greatest influence on GDP in Kazakhstan is made by the magnification factor of oil prices growth rates.
The regression equation for Azerbaijan is as follows: Y=0.8945+ 0.8312x 1 −0.2182x 2 +0.06053x 3 . We tested statistical significance of the equation using the determination coefficient and Fisher's criterion. Overall variability of GDP in Azerbaijan is due to 59.6%-change to x i by R 2 . Tabular value Fkp(3,10)=3.71, Ffact=4.92. When Ffact>Fkp, the determination coefficient is statistically significant and regression equation is statistically reliable. From the highest coefficient β we conclude that the greatest influence on the result of GDP in Azerbaijan is made by the magnification factor of oil production growth rates.
The regression equation for Norway is as follows: Y=0.6723+0 .00859x 1 −0.1098x 2 +0.01839x 3 . We tested statistical significance of the equation using the determination coefficient and Fisher's criterion. Overall variability of GDP in Norway is due to the 13%-change to x i factors by R 2 . Tabular value Fkp (3;10) 3.71, Ffact=0.50. When FFact<Fkp, the determination coefficient is not statistically significant and regression equation is not statistically reliable. From the highest coefficient β we conclude that the greatest influence on the result of GDP in Norway is made by the magnification factor of oil prices growth rates.
See Table 6 for summarized simulation results for the influence of selected factors on economic growth in the countries under consideration.
Note: x1 is 1%-increase in individual structural shift in oil exports share, x2 is 1%-increase in oil production growth rates, and x3 is 1%-increase in oil prices growth rates.
Findings show statistical significance of regression equations in case of Iraq, Libya, Saudi Arabia, Russia, and Azerbaijan. Thus, we generally confirmed our assumption, i.e. the correlation between fluctuations in oil prices and economic growth is stronger in large economies (by absolute value of GDP of oil-exporting countries and hydrocarbons production). We also observed that economic growth in a number of countries largely depends on growth in oil prices and changes to oil production rates. We explain this with the strategy for keeping the balance of income, where the oil price parameter is a tool to choose an ideal production volume. Such strong sensitivity to the price factor does not make it possible to provide for a sustainable export structure (sustainable structure of economy). Therefore, in the long-term, non-resource export development will make it possible to reduce dependence on the influence of oil price fluctuations and achieve sustainable economic growth.

CONCLUSION
Energy resources are of particular strategic importance in global economy development. If at the micro level, we generally talk about an important connective role of energy resources and their involvement in the production process, then, at the macro level, in some countries, exported natural resource is a key source of income, and, consequently, economic growth. From findings we confirm that such an influence grows with economy of scale: larger economies (by absolute figure of GDP in oil-exporting countries and absolute figure of hydrocarbon production) can generate greater economic growth from positively changed oil prices than small economies. We also found the influence of joint coordination policy in oil production on economic growth of nations. The countries under consideration (the OPEC members) might achieve GDP growth from 0.0367% (Iraq) to 0.437% (Libya) in case of 1%-increase in oil production. The OPEC members show the strongest structural shifts in the changed share of oil exports than the other oil-exporting countries under consideration.
Resulting regression dependences show that economic growth of oil-exporting countries largely depends on the very increase in oil prices and changes in oil production rates. This has recently changed behaviours of exporting countries in the oil market as well as the role of the price factor. Maintaining the income balance has become a priority when price shocks are compensated by volumes of oil supply.