Natural Resources Management as Drivers of Economic Growth: Fresh Insights from a Time Series Analysis of Saudi Arabia

Abstract

Natural resources management has played an important role in uplifting the growth performance of countries over the years. However, for the Kingdom of Saudi Arabia (KSA), very little is known regarding the influence of natural resources on economic growth. Therefore, this study focused on investigating the relationship between natural resources and economic growth by focusing on the KSA. This study was based on data for the period 1973–2022 analyzed through “Autoregressive Distributed Lag (ARDL)” modeling. To identify the directions of the relationships between the selected variables, the present study carried out causality testing. The findings indicate that natural resources improved the growth of KSA, which was an indication of the “resource blessing” hypothesis. Other variables, such as education, employment, and investment, also contributed positively to the growth of the KSA economy. Surprisingly, openness to trade decelerated the growth performance. In the short run, again, we found a positive impact of natural resources, education, investment, and employment on growth. Finally, openness to trade maintained its negative impact on growth in the short run. The causality analysis displayed both one-way and two-way relationships between the selected variables. This study suggests that KSA authorities must focus on gearing up the process of economic diversification. Moreover, increased investment, both in physical and human capital, is needed to improve and sustain long-term growth.

1. Introduction

The Kingdom of Saudi Arabia (KSA, hereafter) has abundant natural resources. The natural resources of the KSA include petroleum, natural gas, iron ore, gold, and copper. In 2022, the KSA was the third top exporter of crude oil and was ranked first among the OPEC members, as mentioned by the EIA [1]. To improve the economic growth process, the economy of the KSA has relied heavily on oil rents and total resources rents over the years, as pointed out by Agboola et al. [2]. Like all other resource-rich economies, the KSA has also enjoyed a relatively high per capita income and improved quality of life as compared with the developing world. According to Mati and Rehman [3], the per capita income in the KSA was USD 23,507 in 2021, while the economic growth was 3.2%. Similarly, the human development index (HDI) was 0.875 in 2022, which is an indication of excellent quality of life in the KSA. Among the GCC economies, the HDI value was 0.888 for Bahrain, 0.819 for Oman, 0.875 for Qatar, 0.937 for UAE, and 0.847 for Kuwait in 2022. It is widely accepted that the improved quality of life in the KSA is a reflection of the better economic performance over the years.

The improved economic performance of the KSA is dependent on several factors. Alam and Haque [4] demonstrated that growth in the KSA could be explained by the export sector and domestic investment. However, Alodadi and Benhin [5] showed that exports adversely impacted the growth of the KSA economy. They showed that international tourism arrival and investment are the key determinants of economic growth in the KSA. On the other hand, Bokhari [6] shed light on the important role played by human capital in the growth performance of the KSA economy. Moreover, Aljarallah [7] showed that natural resources improved the productivity and income level in the KSA and endorsed the idea that too much reliance on natural resources is detrimental for sustainable growth. Oil is still the main engine of growth of the KSA, as commented by Albassam [8]. The impact of natural resources on growth is still an open research area, not only globally but also in the case of the economy of the KSA. Still there is no consensus about whether natural resources improve economic growth or impede economic growth, as discussed by Kim and Lin [9]. Studies have largely produced contradictory results. Some studies were optimistic about the “resource blessing” hypothesis, such as Tahir et al. [10], Aljarallah [7], and Kwakwa et al. [11], while others were optimistic about the “resource-curse” hypothesis, such as Zheng et al. [12], Khan et al. [13], and Sachs and Warner [14]. Empirical evidence regarding the resource curse hypothesis is still convincing, as documented by Badeeb et al. [15]. The “resource-curse” theory assumes that the abundance of natural resources harms the growth process. The linkages between resources and growth basically depend on several other policies. These policies include institutional development and human capital development. Poor institutional quality coupled with poor human capital is bound to remove the positive influence of resources on growth. Havranek et al. [16] commented that 40% of published papers have reported no effect and 40% of published papers have reported a negative effect, while 20% of publications have reported a positive relationship between resources and growth. This implies that the natural “resource blessing” and natural “resource-curse” hypotheses are still under research as far as the empirical literature is concerned.

The prime purpose behind this research study was to comprehensively assess the influence of natural resources on the economic growth of the KSA. The KSA is an interesting research case due to its rich resource sector and unique characteristics. The research problem of our study mainly stemmed from the limited research on the resource–growth relationship in the KSA. The economy of the KSA is blessed with abundant natural resources, and these resources could explain the relatively better economic performance over the years. Aljarallah [7] commented that the KSA is the largest economy in the middle east and ranked 18th in the world in terms of GDP. Similarly, the natural gas resources of the KSA are the fifth highest in the world, as pointed out by Mahmood [17]. The empirical literature is indeed very limited on the exact relationship between natural resources and economic growth in the context of the KSA economy. Alkhathlan [18] showed that oil revenues improved the growth of the KSA enormously, while Aljarallah [7] displayed a positive influence of resources on growth. On the other hand, Anser et al. [19] demonstrated that natural gas and minerals had a negative impact, while forest and oil rents positively impacted the growth performance. These observed miscellaneous results in the context of the economy of the KSA on the relationship between natural resources and economic growth were the prime motivations behind the current study.

The novelty of the current research can be seen in several aspects. First, this paper focuses on the relationship between resources and growth, which is no doubt an interesting but little-researched topic. Previous studies on this subject have failed to provide consistent evidence. Second, we contribute to the literature contextually by focusing on the resource-rich economy of the KSA while studying the relationship between natural resources and economic growth. It is noteworthy to mention that the literature on the impact of natural resources on economic growth in the case of the KSA is limited. Third, we were also interested in identifying the directions of relationships between the variables by carrying out causality testing. Therefore, the current study was more comprehensive compared with prior studies and provided fresh evidence on the effects that resources have on growth.

In terms of methodology, the present study focused initially on describing the data in more detail to highlight the trends of variables over the years. In the second step, this study employed cointegration techniques to see the responsiveness of growth to changes in natural resources. Finally, to confirm the directions of the relationships, this study carried out causality testing.

The remaining part of this paper consists of several interconnected sections. The literature review on the determinants of growth is discussed in Section 2. Key statistics on the chosen variables for the KSA are shown in Section 3. The modeling and estimation strategy is presented in Section 4. The results are discussed and analyzed in Section 5. The penultimate section includes a causality analysis, while the final section consists of concluding remarks, implications, and limitations.

2. Literature Review

The relationship between natural resources and growth is a well-debated and well-researched topic. However, there are still disagreements between researchers. Kim and Lin [9] commented that the resource–growth relationship is still a controversial debate due to the inconsistent results reported in the literature. This means that this field of research is still open to further research, as the available literature is inconclusive.

There are mainly two hypotheses. The first hypothesis is known as the “resource-curse” hypothesis, which believes that resource-rich economies have performed poorly in terms of economic growth as compared with countries having little or no resources, as shown by Tahir et al. [10]. The second hypothesis, which is formally known as the “resource blessing” hypothesis, believes that natural resources are a pre-requisite for achieving higher economic growth. The “resource curse” hypothesis received significant attention from researchers and is a well-debated and well-researched area in the literature. The supporters of the natural “resource-curse” hypothesis assume that natural resources are harmful for growth performance. For instance, Gylfason and Zoega [20] endorsed the idea that natural resources negatively impact economic growth and institutional quality. Sachs and Warner [14] demonstrated that natural-resource-abundant countries experienced low growth rates. Kim and Lin [9] also demonstrated that resource-abundant economies experienced relatively poor growth as compared with resource-scarce economies. On the other hand, Gerelmaa and Kotani [21] demonstrated that the recent data do not support the “resource curse” hypothesis. Moreover, Havranek [16] endorsed the idea that there is very weak backing for the “resource curse” hypothesis after controlling for the methodology bias.

The “resource blessing” hypothesis has also received due attention from researchers over the years. For instance, focusing on the experience of the Brunei economy, Tahir et al. [10] showed a positive relationship between resources and growth. Similarly, in the case of GCC economies, Hayat and Tahir [22] validated the resource blessing hypothesis. In the context of the Pakistan economy, Hassan et al. [23] also demonstrated a positive impact that natural resources have on the growth. These findings support the “resource blessing” hypothesis.

In the context of KSA, some recent studies investigated the influence of natural resources on growth. For instance, Aljarallah [7] showed that natural resources improved the productivity and income level in the KSA and endorsed the idea that too much reliance on natural resources is detrimental for sustainable growth. They suggested that the policy of economic diversification is indeed necessary to avoid the “resource curse” hypothesis. Albassam [8] also endorsed the idea that oil is the dominant factor of growth in the KSA. The author further stressed the implementation of economic diversification in the KSA. Research demonstrated that natural gas and minerals had a negative impact, while forest and oil rents positively impacted the growth performance, as shown by Answer et al. [19].

There are still notable disagreements between researchers about the potential role of natural resources. Particularly in the context of the KSA economy, it is pertinent to mention that the research literature is indeed very limited. Hence, the present study was an attempt to fill the mentioned gap by carrying out a comprehensive study on the linkages between natural resources and growth by focusing on the KSA economy.

3. Key Statistics on KSA

In

Table 1. Key statistics.

Variables	1973	1982	% Change	1983	1992	% Change	1993	2002	% Change	2003	2012	% Change	2013	2022	% Change
Y	58,887	36,606	−37.837	33,110	25,205	−23.88	24,379	20,406	−16.296	21,172.9	22,338.5	5.505314	22,629	23,332	3.10743
NRS	67.521	34.97	−48.21	34.089	30.49	−10.57	27.43	38.018	38.574	43.6225	45.5309	4.3748	41.338	22.95	−44.49
INV	8.8341	26.94	204.975	25.157	21.85	−13.13	18.37	18.3	−0.3892	19.1565	23.4672	22.50256	24.927	27.89	11.9028
EDU	1.5599	1.786	14.4768	1.8104	2.060	13.806	2.081	2.3148	11.2115	2.34346	2.58233	10.19314	2.603	2.805	7.7748
OPEN	76.138	86.15	13.1433	79.744	67.806	−14.970	60.07	69.831	16.2471	75.0828	81.9172	9.102446	79.562	62.13	−21.913
EMP	1.8621	3.61	93.8747	4.0096	5.552	38.472	5.555	6.3835	14.9118	6.79037	10.797	59.00465	11.102	15.51	39.7093

Notes: “Authors own analysis using WDI and PWT Data”. “The terms Y, NRS, INV, EDU, OPEN, EMP stands for GDP per capita, natural resources, investment, education, trade openness and employment level respectively”.

, we show key statistics and their behavior during the study period (1973–2022). The decade-wise percentage changes are shown. The income per person decreased during 1972–1982, 1983–1992, and 1993–2002. However, the income per person increased during the last two decades. Similarly, “Natural resources rents (% of GDP)” also decreased during 1973–1982 and 1983–1992. However, an upward trend in “natural resources rents (% of GDP)” was observed during 1993–2002 and 2003 to 2012. Finally, data of the last decades show that “natural resources rents (% of GDP)” decreased. This means that Saudi Arabia is moving in the right direction of economic diversification in recent times.

The domestic investment “Gross fixed capital formation (% of GDP)” showed mixed results. The domestic investment increased enormously in the KSA during 1973–1982. On the other hand, the domestic investment decreased during 1983–1992 and 1993–2002. The data for the last couple of decades demonstrate an upward trend in domestic investment. The human capital “Human capital index, based on years of schooling and returns to education” increased during 1973–2022. The decade-wise data show that human capital increased in all five decades. These improved statistics on education endorsed the overall better performance of the education sector in the KSA.

The “trade (% of GDP)” of the KSA produced mixed results. The statistics show that the “trade openness index” increased during 1973–1982 and decreased during 1983–1992. On the other hand, the trade openness increased during 1993–2002 and 2003–2012. Finally, the trade openness declined during 2013–2022. However, the current trade openness index of the KSA was still above many developing countries. Finally, the employment level improved enormously in the KSA. According to the decade-wise statistics, the employment level “Number of persons engaged (in millions)” increased for all decades. The observed developments in investment, education, and employment could be the key contributors to the growth of the KSA economy over the years.

4. Modeling and Estimations Methods

4.1. The Modeling Strategy

We developed an empirical model to achieve the objective of this study. Previous literature repeatedly documented that economic growth depends on several factors due to its complex and multidimensional nature, as discussed by Tahir and Azid [24]. This implies that in the growth-accounting framework, several factors could be considered. Therefore, besides natural resources, we must consider other factors that could potentially impact economic growth. Domestic investment and human capital have received significant attention from researchers, such as Tahir et al. [10], Barro [25], and Barro [26], due to their enormous growth-promoting benefits. Similarly, liberalized trade policies in the form of increased trade openness were also considered important from the perspective of increased growth, as seen in Dollar [27], Sachs and Warner [14], and Tahir and Azid [24]. Finally, the role of the labor force cannot be ignored in the growth process, as it is the main factor of production. The selected variables were the main determinants of growth in light of the literature. However, some other factors could also be important for growth, including technological progress, government policies, and foreign aid. However, these factors were not included in the current study because of inconsistent data. We specified the following function form for the building of the model for the analysis:

$$\mathrm{E}\mathrm{C}\mathrm{G}=\mathrm{f}({\mathrm{N}\mathrm{R}\mathrm{S}}^{\mathrm{a}},{\mathrm{I}\mathrm{N}\mathrm{V}}^{\mathrm{b}}{,\mathrm{E}\mathrm{D}\mathrm{U}}^{\mathrm{c}}{,\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}^{\mathrm{d}},{\mathrm{E}\mathrm{M}\mathrm{P}}^{\mathrm{e}})$$

(1)

Model 1 indicates that economic growth is dependent on natural resources, investment, education, openness to trade, and employment level. Assuming non-linearities, we transformed model 1 as given below:

$${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}+{{\mathsf{\beta }}_2\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}+{{\mathsf{\beta }}_3\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}+{{\mathsf{\beta }}_4\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}+{{\mathsf{\beta }}_4\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}+{\mathrm{U}}_{\mathrm{t}}$$

(2)

In model 2, the economic growth was captured by taking the “growth of GDP per capita” in real terms. For approximating the natural resources, the “total natural resources rent (% of GDP)” was used. This measure of natural resources was supported by previous studies [10,22]. The domestic investment in the economy was measured by taking the “gross fixed capital formation (% of GDP)”, while for the human capital, the “Human capital index, based on years of schooling and returns to education” was used. Finally, for the trade openness, the “total trade (% of GDP)” was utilized, while for the employment level, the “Number of persons engaged (in millions)” was used. The data were taken from the WDI and Penn World Tables for the period 1973–2022.

4.2. Estimation Methods

Researchers mentioned that the unit root may be linked with time series data, as pointed out by Tahir [28] and Tahir et al. [29]. It basically violates the assumption of a well-known estimator: ordinary least squares (OLS). In such a situation, researchers suggested using cointegration analysis instead of using the traditional OLS method. Cointegration is defined as the “stationary relationship that may exist between non-stationary variables”, as shown by Tahir [28]. Several cointegration tests are available in the literature. The Engle and Graner [30] bivariate cointegration test works well for two non-stationary variables. However, the two-variable case is quite rare in applied research studies. Keeping this fact in mind, Johansen [31] proposed the multivariable cointegration testing framework, which is relevant for models with multiple variables. The pre-requisite for the cointegration approaches is that all variables should possess the unit root problem at the level. Pesaran et al. [32] proposed the “Autoregressive Distributed Lag (ARDL, hereafter)”, which can take care of variables with different orders of integration.

4.3. ARDL Modeling

There are several notable advantages linked with ARDL modeling. ARDL modeling works efficiently with a small sample, as endorsed by Tahir [28]. Similarly, ARDL modeling can take care of variables with a diverse order of integration. We converted model 2 into the ARDL approach, as given below:

$$\begin{array}{c}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_2\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_3\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-1}{+{\mathsf{\lambda }}_4\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_5\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_6\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(3)

$$\begin{array}{c}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_2\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_3\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-1}{+{\mathsf{\lambda }}_4\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_5\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_6\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(4)

$$\begin{array}{c}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_2\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_3\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-1}{+{\mathsf{\lambda }}_4\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_5\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_6\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(5)

$$\begin{array}{c}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_2\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_3\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-1}{+{\mathsf{\lambda }}_4\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_5\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_6\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(6)

$$\begin{array}{c}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_2\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_3\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-1}{+{\mathsf{\lambda }}_4\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_5\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_6\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(7)

$$\begin{array}{c}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_2\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_3\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-1}{+{\mathsf{\lambda }}_4\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_5\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-1}+{{\mathsf{\lambda }}_6\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(8)

Expressions (3)–(8) are the ARDL representations of model 2. In all the models, the parameters (${\mathsf{\beta }}_1–{\mathsf{\beta }}_6)$ capture the short-run relationships. On the other hand, the parameters (${\mathsf{\delta }}_1$–${\mathsf{\delta }}_6)$ stand for the long-run relationships. The hypotheses provided below were tested to accept or reject the cointegration:

$${\mathrm{N}\mathrm{u}\mathrm{l}\mathrm{l}\mathrm{h}\mathrm{y}\mathrm{p}\mathrm{o}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{s}:\mathrm{“}(\mathsf{\lambda }}_1={\mathsf{\lambda }}_2={\mathsf{\lambda }}_3={\mathsf{\lambda }}_4={\mathsf{\lambda }}_5={\mathsf{\lambda }}_6=0)\mathrm{”}$$

$$\mathrm{A}\mathrm{l}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{h}\mathrm{y}\mathrm{p}\mathrm{o}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{s}:\mathrm{“}({\mathsf{\lambda }}_1\ne {\mathsf{\lambda }}_2\ne {\mathsf{\lambda }}_3\ne {\mathsf{\lambda }}_4\ne {\mathsf{\lambda }}_5\ne {\mathsf{\lambda }}_6\ne 0)\mathrm{”}$$

5. Results and Analysis

5.1. Unit Root Results

The unit root testing was performed using the “Augmented Dickey Fuller (ADF)” and “Phillips-Perron (PP)” tests.

Table 2. Unit root testing.

“Variables”	“ADF Test”		“PP Test”
	“Level”	“Diff.”	“Level”	“Diff.”	“Conclusion”
${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$	−1.259	−3.422 **	−1.166	−5.041 ***	I(1)
${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$	−2.878	−6.544 ***	−2.910	−8.544 ***	I(1)
${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$	−3.372 *	−8.446 ***	−5.445 ***	−16.475 ***	I(0)
${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$	−2.003	−9.201 ***	−0.060	−3.614 **	I(1)
${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$	−3.091	−9.716 ***	−2.196	−7.009 ***	I(1)
${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	−2.956	−5.240 ***	−2.789	−5.289 ***	I(1)

Notes: “The terms ECG, NRS, INV, EDU, OPEN, EMP stands for economic growth, natural resources, investment, education, trade openness and employment”. “The asterisk (***), (**) and (*) shows significance at 1, 5 and 10 percent level respectively”.

includes the results. At the level, the results confirmed the presence of unit root testing with all variables except domestic investment. However, the unit root problem disappeared at the first difference. The variables chosen for this study possessed a non-uniform order. Hence, the ARDL analysis was appropriate.

5.2. The Bound Test

The ARDL findings are depicted in

Table 3. Bound test findings.

Specifications	F Test	Cointegration
${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}/{\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$, ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$, ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	4.797	Yes
${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}/{\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$, ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$, ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	5.460	Yes
${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}/{\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$, ${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$, ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	5.259	Yes
${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}/{\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$, ${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$, ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$, ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	41.283	Yes
${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}/{\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$, ${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$, ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	3.607	Yes
${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}/{\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$, ${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$, ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$, ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}},$ ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$	3.383	Yes
Critical level (%)	L.B	U.B
1	3.06	4.15
5	2.39	3.38
10	2.08	3.00

Note: “The terms ECG, NRS, INV, EDU, OPEN, EMP stands for GDP per capita, natural resources, investment, education, trade openness and employment”.

. The lower part of Table 3 includes the critical values. Cointegration was accepted, as the F-test values were higher than the critical values for all the estimated models. Therefore, we concluded that there was a cointegrating relationship between the variables.

The bound test provided sound evidence to accept the cointegrating relationship. The second step in the ARDL modeling was to develop the “error correction models (ECM, hereafter)”. The ECMs have two purposes. The first purpose of the ECMs was to investigate the short-run impacts of the independent variables on the dependent variable. The second purpose of the ECMs was to figure out the adjustment speed of the models. The ECM models are expressed below:

$$\begin{array}{c}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}={\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{T}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(9)

$$\begin{array}{c}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{T}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(10)

$$\begin{array}{c}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{T}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(11)

$$\begin{array}{c}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{T}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(12)

$$\begin{array}{c}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{T}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(13)

$$\begin{array}{c}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}{=\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}1}{\mathsf{\beta }}_{1\mathrm{i}}{∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}2}{\mathsf{\beta }}_{2\mathrm{i}}{∆\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}3}{\mathsf{\beta }}_{3\mathrm{i}}{∆\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}4}{\mathsf{\beta }}_{4\mathrm{i}}{∆\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=0}^{\mathrm{n}5}{\mathsf{\beta }}_{5\mathrm{i}}{∆\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=0}^{\mathrm{n}6}{\mathsf{\beta }}_{6\mathrm{i}}{∆\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}-\mathrm{i}}+{{\mathsf{\lambda }}_1\mathrm{E}\mathrm{C}\mathrm{T}}_{\mathrm{t}-1}+{\mathrm{e}}_{\mathrm{t}}\end{array}$$

(14)

5.3. Analysis of Descriptive Statistics

The descriptive statistics are displayed in

Table 4. Descriptive statistics.

	${\mathbf{Y}}_{\mathbf{t}}$	${\mathbf{N}\mathbf{R}\mathbf{S}}_{\mathbf{t}}$	${\mathbf{I}\mathbf{N}\mathbf{V}}_{\mathbf{t}}$	${\mathbf{O}\mathbf{P}\mathbf{E}\mathbf{N}}_{\mathbf{t}}$	${\mathbf{E}\mathbf{D}\mathbf{U}}_{\mathbf{t}}$	${\mathbf{E}\mathbf{M}\mathbf{P}}_{\mathbf{t}}$
Mean	21,094.96	38.677	21.499	75.307	2.165	6.741
Maximum	35,689.59	87.284	29.990	120.619	2.713	13.744
Minimum	15,560.48	17.318	8.834	49.713	1.553	1.857
Std. dev.	6060.998	14.034	3.755	13.887	0.372	3.555
Observations	50	50	50	50	50	50

. The mean value of the GDP per capita was USD 21,094.960 in real terms. The maximum value of the GDP per capita (USD 35,689.590) was observed in 1974, while the minimum value (USD 15,560.480) was recorded in 1987. Similarly, the average value of the natural resource rents (% of GDP) was 38.677%, with a notable standard deviation of 17.318%. The highest contribution of natural resources (87.284%) was experienced by the KSA economy in 1979, while the lowest contribution (17.318%) was observed in 2020. Similarly, the contribution of investment in GDP was 21.499%. The maximum (29.990%) and minimum (8.834%) contributions of investment in the GDP were recorded in 1979 and 1974, respectively.

The mean value of trade during the study period was 75.307%, which is impressive. The maximum statistic of trade was 120.619%, which was recorded in 1973, while the minimum statistic of trade (49.713%) was recorded in 2020. The average value of the “human capital index, based on years of schooling and returns to education” was 2.165, with a standard deviation of 0.372. The maximum (2.713) and minimum (1.553) values of human capital were observed in 2019 and 1973 respectively. Finally, the employment level “Number of persons engaged (in millions)” had an average value of 6.741.

5.4. Long-Run Results

Table 5. Long run results.

Variables	Coefficient	S.D.E	Prob. V
${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$	0.319 ***	0.073	0.000
${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$	0.387 **	0.154	0.018
${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$	1.864 **	0.763	0.022
${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$	−0.688 ***	0.186	0.001
${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	0.360 ***	0.111	0.003
Constant	−0.869	0.656	0.197

Notes: “The terms NRS, INV, EDU, OPEN, EMP stands for natural resources, investment, education, trade openness and employment”. “(***): 1%, (**): 5%”, “S.D.E: Standard Error”.

includes the long-run results. According to the results, natural resources influenced the growth performance of the KSA positively. The point estimate indicates that a one percent rise in natural resources enhanced the growth by 0.319 percent. The observed findings were the indication of the presence of “resource blessing” hypothesis in the KSA. In other words, the traditional “resource curse” hypothesis was not valid in the KSA. Previous studies, such as Shittu [33] and Tahir et al. [29], also provided evidence about the presence of the “resource blessing” hypothesis. On the other hand, our results differ from the results demonstrated by some researchers, such as Yu [34] and Rahim et al. [35]. The KSA authorities should improve the process of economic diversification. Relying on natural resources for long-term growth is not a rational policy.

Domestic investment accelerated the growth process of the KSA significantly, as its coefficient was positive and significant in the estimated models. Domestic investment is considered as an engine of economic growth according to previous studies, such as Tahir [24] and Tahir [36]. Our findings regarding investment and economic growth are supported by prior literature, such as Shabbir [37], Bakari [38], and Tahir [24]. Similarly, human capital also impacted the growth positively in the KSA. Our results are consistent with recent research studies, such as Maneejuk and Yamaka [39] and Kotásková et al. [40]. Therefore, the KSA policymakers need to pay full attention to investing in education.

The results indicate that the employment level contributed positively to the growth process of the KSA. The observed positive links between employment and economic growth were expected, as employed individuals contribute to the production process as compared with the unemployed. Thus, providing more employment opportunities for the labor force will help the KSA economy to improve the growth process further.

The results regarding trade and growth came as a surprise. Our results suggest that economic growth negatively responded to increased trade openness. The beneficial influence of trade on growth was dependent on other factors, such as inflation and financial development. Therefore, the KSA authorities are suggested to focus on developing their financial sector and ensure stability in prices to enjoy the potential positive impacts of trade openness. The statistics presented in Table 1 show that the openness degree of the KSA economy decreased significantly over the years, which could be one of the reasons behind the negative influence of trade on growth.

5.5. Discussion on Short-Run Results

In the short run, we found positive relationships between natural resources, investment, education, employment, and economic growth as mentioned in

Table 6. Short-run results.

“Variables”	“Coefficient”	“S.D.E”	“Prob. V”
$∆{\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$	0.225 ***	0.055	0.000
$∆{\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$	0.249 **	0.117	0.043
$∆{\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$	2.598 **	1.101	0.026
$∆{\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$	−0.231 *	0.131	0.090
${∆\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	0.792 ***	0.233	0.002
ECT (−1)	−0.489 ***	0.108	0.000

Notes: “The terms NRS, INV, EDU, OPEN and EMP, ECT stands for natural resources, investment, education, trade openness, employment and error correction term”. “(***): 1%, (**): 5%, (*): 10%”, “S.D.E: Standard Error”.

. Similarly, a negative relationship was observed between trade and growth. Therefore, the short-run results validated the long-run findings. In other words, the variables selected for the current study were the robust determinants of economic growth. Hence, these results could be used for policy analysis by the KSA authorities. Finally, the estimated model had a strong adjustment speed of 48%, which is excellent.

5.6. Diagnostic Analysis

Reliability analysis is considered important in applied research. Keeping this fact in mind, the present study conducted several tests as shown in

Table 7. Diagnostics analysis.

“Test”	“Values”	“Conclusion”
“LM”	1.442 (0.257)	“No serial correlation”
“JB”	0.970 (0.615)	“Data is normal”
“Ramsey”	2.487 (0.105)	“Correct functional form”
“White”	2.552 (0.091)	“No heteroscedasticity”

Notes: “LM: Lagrange Multiplier”, “JB: Jarque-Bera”.

in the light of prior literature [22,28,29]. The LM test confirmed that our estimated models were free from the undesirable serial correlation problem. Similarly, the data used in the analysis had a normal structure, as was evident from the Jarque–Bera test. Furthermore, the absence of heteroscedasticity was validated by the White test, while the Ramsey test highlighted the proper functional form used.

5.7. Stability Testing

We also conducted the CUSUM test and its square. The graphical presentation of the tests is shown in Figure 1 and Figure 2. Both figures indicate stability as the blue lines are lying inside the critical lines.

6. Causality Analysis

The causality testing was performed to highlight the direction of each relationship (

Table 8. Causality testing.

“Directions”	“F-Test”	“Prob”
${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$ $\to {\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$	0.357	0.701
${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$ $\to$ ${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$	1.011	0.372
${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$	1.654	0.203
${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$ $\to$ ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$	4.804 ***	0.013
${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$	2.679 *	0.080
${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$ $\to$ ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$	4.528 **	0.016
${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$	0.170	0.843
${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$	1.688	0.196
${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$	0.144	0.866
${\mathrm{E}\mathrm{C}\mathrm{G}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	0.681	0.511
${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$ $\to$ ${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$	1.883	0.164
${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$ $\to$ ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$	4.271 **	0.020
${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$ $\to$ ${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$	1.396	0.258
${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$ $\to$ ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$	5.359 ***	0.008
${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$ $\to$ ${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$	0.675	0.514
${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$	0.126	0.881
${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$ $\to$ ${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$	1.035	0.363
${\mathrm{N}\mathrm{R}\mathrm{S}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	0.697	0.503
${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$ $\to$ ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$	5.624 ***	0.006
${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$ $\to$ ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$	3.978 **	0.026
${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$ $\to$ ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$	0.797	0.456
${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$	0.177	0.837
${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$ $\to$ ${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$	1.023	0.367
${\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	1.345	0.271
${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$ $\to$ ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$	0.329	0.721
${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$	1.242	0.298
${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$ $\to$ ${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$	0.867	0.427
${\mathrm{O}\mathrm{P}\mathrm{E}\mathrm{N}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	1.905	0.161
${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$	5.227 ***	0.009
${\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{t}}$ $\to$ ${\mathrm{E}\mathrm{M}\mathrm{P}}_{\mathrm{t}}$	1.34713	0.270

Note: “(***): 1%, (**): 5%, (*): 10%”.

). According to the results obtained, openness and growth were bidirectionally connected. Similarly, investment and trade openness were also connected with each other bidirectionally. Growth was unilaterally related to investment, while natural resources were unidirectionally linked with investment and trade openness. Finally, a one-way relationship that ran from employment level toward education was observed.

7. Conclusions and Implications

7.1. Conclusions

This research paper examined the influence of natural resources on growth by utilizing the data of the KSA. This study used data from 1973 to 2022 and employed the ARDL approach of cointegration and causality techniques.

Our results show that the long-term growth of the KSA was impacted positively by the natural resource sector. The rejection of the “resource curs” hypothesis was the acceptance of the “resource blessing” hypothesis. Other than natural resources, both education and investment also had positive influences on growth. Similarly, the employment level positively impacted growth, while trade negatively impacted growth. Natural resources, investment, employment, and education appeared to be the main determinants of growth in the short run. Moreover, the negative impact of trade on growth also did not alter. The causality analysis showed that trade was bidirectionally related to both growth and investment. Moreover, economic growth was unilaterally related to investment, while natural resources were unidirectionally linked with investment and trade openness. Finally, a one-way relationship that ran from employment level toward education was observed.

7.2. Implications

This study suggests that the KSA economy should embrace economic diversification despite the observed positive relationship. Relying on natural resources for long-term growth is not a rational policy. Similarly, the KSA authorities are suggested to take visible steps to invest more into physical capital, as well as in human capital. Both these factors will improve the growth process. Moreover, increased domestic investment is also needed for the employment of the labor. Finally, the existing trade policies must be re-evaluated.

7.3. Future Research Directions

Based on the findings, we recommend that potential researchers should validate the models used in this study in some other contexts to ensure their reliability. Second, the present study recommends that potential researchers should include some more control variables to establish a robust relationship. Lastly, our study recommends that future researchers should utilize some more advanced econometric tools to provide more in-depth analysis.