Global Supply Chain Distribution and Natural Resources in the Era of Digitalization

Abstract

This study examines the effects of supply chain disruptions and ICT product exports on natural resource rents in European countries between 2004 and 2022. The findings show that strong supply chains increase natural resource rents, while ICT product exports support environmental sustainability and reduce natural resource rents. Patents reduce natural resource rents, while investments made by financial institutions in resource-intensive sectors increase natural resource rents. In addition, urban population growth was found to put pressure on natural resources, leading to a decrease in natural resource rents. Importantly, economic growth has no significant effect on natural resource rents in OECD countries.

1. Introduction

In the 21st century, one of the decisive factors in the reconstruction of the global economy has been digitalization. Especially with the spread of information and communication technology (ICT), production, logistics, service, and foreign trade processes have been transferred to digital fields, bringing change and transformation to global supply chains. In addition to providing operational productivity, digitalization has become decisive in terms of natural resource use, environmental sustainability, and energy efficiency. Therefore, revealing the environmental impact of the digitalization of foreign trade is of strategic importance in development policies and international economies, as well as environmentally.

Aspects of traditional supply chains, such as excessive numbers of vehicles, inadequate data sharing, and high logistics costs, are not usually considered in economic and environmental terms. Supply chain distribution (SCD) is a comprehensive process that expresses the delivery of raw material to the final consumer. The size of this process and the route involved, as well as the financial situation, all directly change the functioning of the system. The concept of “Intangible Flow Theory”, developed by [1,2] for flows of physical goods, states that having the same rate is important for information supply chains of information and financial flows. Thanks to digital technologies, the management of non-concrete workflows can be improved, and thus, the environmental performance in all rings of the supply chains can be improved [3,4].

Recently, studies have been carried out on environmental outputs in such fields as green logistics, energy efficiency, and resource management [5,6,7]. However, these studies generally consider micro-level firm data or ICT usage rates throughout a country; they do not adequately consider the impact of foreign trade on the natural resource rent (NRR) of ICT and SCD exports from digital components. In addition, ICT’s environmental output should not be considered only from a positive angle. For example, Hilty and Aebischer [8] have drawn attention to negative environmental outputs such as increasing energy consumption and e-time formation in the production processes of ICT infrastructure. Therefore, when considering the impact of digitalization on natural resources, it is necessary to deal with these multidimensional and different angles.

In the present study’s modeling approach, it was assumed that the effects of ICT and SCD exports on NRR may not be linear. Therefore, by going beyond the classical fixed effects models, the KRLS (kernel-based regularized least squares) method, which can capture non-linear relationships and model the marginal effects of variables more flexibly, was preferred. Thus, the potential threshold effects and heterogeneous distribution characteristics of digital foreign trade on natural resource use could be analyzed more accurately.

In this context, the focus of the current study is to empirically examine the impact of ICT and SCD exports on natural resource rent in European countries. The ability to distinguish the effects of digitalization on physical (goods) and abstract (knowledge and finance) flows is not only about environmental impacts, but also concerns the sustainability of global trade. With this study direction, the research question of the study was determined as follows:

“What is the effect of European countries on the natural resource rent of European countries in 2000–2022?”

To answer this research question, the effects of ICT and SCD exports on natural resource use were examined by comparative panel data methods. While ICT points to information-based foreign trade products, SCD represents export categories based on supply chain software, hardware, and system solutions. This difference allows the effects of structural components of digital foreign trade on natural resource use to be more clearly determined. It also has the potential to make the indirect effects of digital transformation on national resource policies visible.

The current study has three contributions. First, it is one of the first comprehensive data studies to adopt an approach that takes into account foreign trade channels and the effects of ICT and SCD exports on natural resource rent. Secondly, it makes an original contribution to the literature by considering the impact of digital foreign trade in terms of environmental sustainability from a theoretical point of view within the framework of “Intangible Flow Theory”. Third, within a framework of evidence from European countries, it offers guidance to policymakers on how ICT-based exports can affect resource efficiency and environmental pressures. In this respect, the study aims to present original and applicable results for stakeholders who have to make important decisions in the areas of digitalization, foreign trade, and sustainability.

Considering all these, digitalization will bring benefits and challenges in the field of global SCD and in the natural resources (NRs) market. Effectively integrating ICT into the global supply chain is expected to result in improved resource management and more efficient supply chains. However, researchers should also comprehensively evaluate the environmental impact of digital technology. This research article examines the complexities of these processes and addresses the impact of SCD and ICT on NRs. For this purpose, we aimed to make inferences through analyses using empirical data obtained from Europe for the period 2000–2022.

2. Literature Review

The relationship and developments among supply chains, information, ICT, and NRs have significant socio-economic, technological, and environmental impacts. This section includes a literature review on the effects of supply chains and ICT on NRs. It summarizes both the risks and positive effects of developments in these areas, particularly in terms of environmental pollution.

2.1. Impact of Supply Chain Distribution (SCD) on Natural Resources (NRs)

Studies on the interaction between SCD and NRs have been increasing in recent years. Due to increasing globalization and consumer expectations in the growing world economy, the efficiency of supply chains and not increasing environmental risks have become very important. This literature review brings together current findings on the impact of SCD on NRs, new methods, and results that have emerged to promote resource depletion, environmental degradation, and sustainable management.

SCD can also have a negative impact on the consumption of NRs. Since the supply, refining, and transportation of raw materials require significant energy inputs, this results in the depletion of non-renewable resources. The transportation sector, one of the most important components of supply chains, uses 30% of the world’s energy obtained from fossil fuels [9]. Excessive consumption of fossil fuels and other NRs increases dependence on non-renewable resources [10].

As mentioned above, global SCD directly contributes to environmental degradation. Various sectors, including transportation and industrial processes, significantly increase greenhouse gas emissions, thereby accelerating climate change and worsening air pollution. In 2018, carbon dioxide emissions from freight transportation alone accounted for 8% of total global emissions [11]. Furthermore, supply chain transportation, especially in sectors such as agriculture and manufacturing, leads to significant reductions in land, forest, and biodiversity [12].

Supply chain operations also have a significant impact on water resources [13]. Transportation and manufacturing activities cannot be ignored in terms of increasing water consumption and pollution. For example, the textile industry supply chain consumes significant amounts of water and discharges toxic substances into water resources [14]. Moreover, the increasing agricultural supply chain also uses significant amounts of water resources for irrigation and therefore can cause water scarcity in certain areas [13].

Another important component of supply chain distribution is waste generation. Supply chain inefficiencies often lead to excessive waste, which can put a strain on natural resources. For example, packaging waste is a major problem, with millions of tons of plastic packaging waste polluting the environment every year [15]. The implementation of efficient waste management strategies, such as recycling and reuse, is crucial to minimizing the negative environmental problems caused by waste. However, according to the literature, current techniques are inadequate and are often implemented below standards [16].

Although SCD has negative consequences, it also presents an opportunity to conserve NRs. Here, it is necessary to pay attention to what can be done to conserve NRs. Adopting green supply chain management (GSCM) strategies such as incorporating renewable energy, using environmentally friendly packaging, and optimizing transportation can effectively minimize environmental consequences [17]. Initiatives such as the EU’s Green EU Consensus, as well as regional cooperation and trade agreements, reduce negative environmental impacts. In addition, some global companies are supporting these efforts by making policy changes to increase the sustainability of their supply chains. These are seen as a necessity to increase energy efficiency globally and reduce negative environmental outcomes [18].

2.2. The Impact of ICT on Natural Resources (NRs)

The impact of NRs on environmental quality is related to ICT development. This is because the production and management technologies that ICT has introduced have prepared the environment for more efficient use of NRs [19]. In addition, ICT contributes to the transformation of resource uses that produce environmentally harmful outputs into environmentally friendly uses [20]. Low-tech extraction methods, especially those related to the consumption of oil and coal, which are non-renewable energy sources, generally increase soil and water pollution. Fossil substances must be processed before they can be used. Polluting gases and wastewater emerge after this process. ICT, consisting of mobile networks, smart computing, and big data applications, provides a more environmentally friendly production and supply chain [21]. In addition, the development of ICT has reduced the cost of clean energy and increased the demand for renewable energy use, thus increasing the share of renewable energy use [22]. Increasing technology and new developments reduce the cost of renewable natural resources and enable the transition to more efficient industries [23]. There is a consensus that using renewable energy has an important advantage in terms of ensuring environmental sustainability [24,25]. However, if ICT is not developed, how useful can using NRs and clean energy be?

In addition, ICT applications can improve environmental quality by reducing the negative impact of NRs on environmental quality. Improving the applications used in industry can reduce environmental pollution by contributing to the development of cleaner energy sources and reducing transportation emissions [26]. Knowledge-based industries brought by ICT can provide opportunities for resource-dependent countries to diversify their economies, thus reducing resource use and environmental risks, which are indirect environmental effects [27]. With the development of the internet and communication tools, ICT contributes to increasing environmental awareness. In addition, social media and mobile applications increase the flow of information and interaction between the public and citizens, which helps governments pursue more environmentally friendly policies and prevent corruption [28]. Moreover, ICT contributes to open communication between resource users and the public, thus paving the way for more effective and efficient management of NRs [29]. All these developments are indirect developments that reduce the negative effects of NR use. In a study conducted for 146 economies, ICT led to a decrease in global CO₂ emissions [30]. In a study conducted in China, ICT contributed to the renewable energy transition in China [31]. Another study conducted for the BRICS countries between 2000 and 2017 shows that ICT products are energy-efficient and reduce CO₂ emissions while encouraging renewable energy consumption [32]. In this context, when existing studies are considered, it can be said that the level of development in ICT can determine the impact of natural resources on the environment. In other words, it is important to take ICT into account when discussing the link between NRs and environmental quality. In addition, although some studies have investigated the role of ICT in natural resource management or environmental protection, there is a research gap in how ICT regulates the relationship between natural resources and environmental quality.

3. Data and Methodology

3.1. Data

The main objective of this study is to analyze the impact of SCD and ICT goods exports on NRR for European countries. In addition to this main objective, the impact of RGDPC, URB, FIE, and PATN on NRR is analyzed for the period 2004–2022. The reason for choosing European countries in the study is that Europe serves as a unique laboratory for this study due to its different economic structures, advanced digital environment, and strict environmental regulations.

The dependent variable of the study is total natural resource rent (NRR). NRR refers to the income or profit from the extraction and utilization of NRs. It is defined as the income earned by private companies, governments, or individuals from the utilization of NRs in excess of the costs of extraction, production, and distribution [6]. These revenues can be obtained from various sources such as minerals (such as oil, gas, and coal), forest products, fisheries, and agriculture [33]. In addition to contributing to economic growth, NRR can also cause problems such as income inequality and environmental destruction. Therefore, effective management and fair distribution of NRs are important for sustainable development and regional prosperity [34]. Figure 1 shows the average NRR scores of European countries. Accordingly, Norway has the highest NRR.

Supply chain distribution (SCD) is one of the important factors affecting natural resource utilization. Disruptions in SCD can create difficulties in the supply of raw materials, increasing demand and leading to overexploitation of rare resources. In this context, a strong SCD is expected to support NRR [35]. RGDPC and FIE are particularly important for the extraction of natural resources. High RGDPC and FIE increase NRR by enabling the use of new technologies for natural resource extraction.

Population growth affects the utilization of natural resources and thus the NRR rent. The increasing population may increase the demand for and dependence on natural resources. Such an effect can lead to more intensive use of resources and increased demand. As a result, the rent from natural resources may decrease. Patents may also indirectly affect the rent from NRs. New technologies and innovations can enable more efficient use of NRs and make the extraction or processing of resources less intensive. This can contribute to the sustainable management of natural resources and a more equitable distribution of natural resource rent. The export of information and ICT products and materials generally has a natural-resource-rent-reducing effect. This is because ICT products are generally knowledge- and technology-intensive and have lower natural resource utilization. Information on the variables is given in

Table 1. Information about variables.

Abbreviation	Calculation	Source
NRR	Total natural resources rent (% of GDP)	World Bank
SCD	Supply chain distribution	FED: https://www.newyorkfed.org/ (accessed on 20 April 2025)
ICT	ICT goods exports (% of total goods exports)	World Bank
PATN	Applicants for patents—total number	Eurostat
FIE	Financial institutions efficiency	IMF
URB	Urban population (% of total population)	World Bank
RGDPC	$\mathrm{Growth} \mathrm{Rate} =\left({\displaystyle \frac{{\mathrm{RGDPC}}_t-{\mathrm{RGDPC}}_{t-1}}{{\mathrm{RGDPC}}_{t-1}}}\right)\times 100$	Eurostat

.

3.2. Methodology

To mitigate potential selection and reporting biases, variable definitions and sources were standardized and verified across all data providers. No imputation was applied for missing values. Since the study relied solely on publicly available, de-identified macro-level data, no ethical approval was required.

In this study, we utilized the kernel-based regularized least squares (KRLS) method, recently introduced by [36], as our primary analytical tool. KRLS is a robust estimation technique grounded in machine learning algorithms, capable of capturing intricate relationships [37]. Beyond merely identifying relationships, this method facilitates the estimation of the marginal effects of explanatory variables. Unlike the least squares (LSLS) method, which directly models the connection between sample points and target values, the KRLS estimator employs a kernel to link inputs and target values. This provides a flexible regression framework that streamlines computations, utilizing non-linear and non-summative functions of variables [36]. This characteristic allows for the demonstration that the impacts on the dependent variable can vary over specific time intervals. The econometric specification of the kernel-based regularized least squares (KRLS) model can be expressed as follows:

$${\mathit{Y}}_{\mathit{i}}=\mathit{f}\left({\mathit{X}}_{\mathit{i}}\right)+{\mathit{ϵ}}_{\mathit{i}}$$

The equation represents a model where Y_i is the dependent variable for observation i, X_i is a vector of explanatory variables for observation i, f( ) is an unknown function to be estimated, and ϵ_i is the stochastic error term.

Unlike parametric methods that assume a specific functional form for f(⋅), KRLS estimates this function flexibly using a kernel-based approach. The model approximates f(⋅) as a linear combination of kernel functions evaluated at each data point:

$$\mathit{f}\left(\mathit{X}\right)=\sum _{\mathit{i}=1}^{\mathit{N}}{\mathit{c}}_{\mathit{i}}\mathit{K}\left({\mathit{X}}_{\mathit{i}},\mathit{X}\right)$$

where K(X_i, X_j) is a positive semi-definite kernel function that measures the similarity between observations X_i and X_j, and c_i represents weights to be estimated.

K(X_i, X_j) captures the similarity between two data points, and KRLS constructs a non-linear model using this similarity measure. The coefficient c_i represents the contribution of each data point to the model. These coefficients are learned to minimize the error between the predicted values and the observed values.

As a robustness check, we opted for the bootstrap quantile regression (BSQREG) estimator. Defined by Hahn [38], the bootstrap quantile regression method aims to estimate both the target and explanatory quantiles. It operates by repeatedly resampling the sample data, subsequently estimating the coefficients of determination and their significance. This resampling technique determines the model’s coefficients by selectively retaining certain data points while omitting others. Both bootstrap and quantile regression methods are powerful statistical tools for constructing confidence intervals and conducting hypothesis tests. A key advantage of these methods is their ability to resample data, which allows for statistical inference without the assumption of asymptotic normality in the samples. Consequently, they are frequently employed in practical scenarios involving small sample sizes or when the underlying data distribution is unknown or non-normal [39]. Bootstrap is a technique used to better understand statistical uncertainties by resampling from the original dataset. The BSQREG method combines these two techniques. The econometric specification of the BSQREG model can be expressed as follows:

Quantile regression is used to model the conditional distribution at a particular quantile (such as the median or 90% quantile). We can express the model as follows:

$$Q_{\mathsf{\tau }}\left(Y∣X\right)=X_i{\mathsf{\beta }}_{\mathsf{\tau }}$$

This formula is the basic expression of quantile regression. Here, Q_τ(Y∣X) represents the estimate of the τ-th quantile of Y_i, and X_iβ_τ represents the estimated value obtained by multiplying the regression coefficients of the independent variables. The main advantage of quantile regression is that it can model relationships across specific quantiles, not just the mean.

The basic model of the study is as follows:

$$NR{R}_{it}=\mathsf{\alpha }+{\mathsf{\beta }}_{1}SC{D}_{it}+{\mathsf{\beta }}_{2}IC{T}_{it}+{\mathsf{\beta }}_{3}RGDP{C}_{it}+{\mathsf{\beta }}_{4}UR{B}_{it}+{\mathsf{\beta }}_{5}FI{E}_{it}+{\mathsf{\beta }}_{6}PAT{N}_{it}+{\mathsf{ϵ}}_{it}$$

This baseline panel regression model is constructed to investigate the determinants of natural resource rent (NRR) in European countries for the period 2004–2022. The dependent variable is natural resource rent as a percentage of GDP. The model includes supply chain distribution (SCD), ICT product exports (ICT), real GDP per capita (RGDPC), urban population rate (URB), financial institutions’ efficiency (FIE), and the number of patent applications (PATN) as explanatory variables. Based on empirical findings, the model captures how supply chain dynamics and digital trade influence environmental outcomes through their impact on natural resource rents, while controlling for economic growth, financial activity, technological innovation, and urbanization.

4. Findings and Discussion

4.1. Findings

In this section, the findings of the econometric analyses are reported, interpreted, and discussed.

Table 2. Descriptive statistics.

Variable	Mean	Std. Dev.	Min	Max	VIF
NRR	0.783	1.602	0.000	12.617	—
RGDPC	1.980	4.039	−14.5	23.300	1.11
SCD	0.110	0.656	−0.725	1.950	1.00
ICT	6.519	6.968	0.067	51.126	1.11
PATN	5.782	2.267	0.000	10.215	1.30
FIE	0.578	0.081	0.200	0.830	1.12
URB	73.729	12.881	51.308	98.153	1.20

contains descriptive information about the variables and the VIF value.

VIF values range from 1.00 to 1.30, and there is no concern regarding multicollinearity for all variables. This result shows that the variables in the model are not highly correlated with each other.

Figure 2 contains the correlation matrix.

Table 3. CDS result.

NRR	33.11	0.000
RGDPC	61.23	0.000
SCD	87.83	0.000
ICT	14.89	0.000
LNPATN	46.52	0.000
FIE	8.10	0.000
URB	54.62	0.000

presents the results of the cross-sectional dependence test. CSD tests whether the effect of a shock occurring in one country is observed in other countries.

When the results of Figure 2 are analyzed, it is determined that all variables contain horizontal cross-section dependence. Since the variables contain CSD, the stationarity of the variables is examined with the Pesaran [40] CADF test, which takes CSD into account in the following unit root test and is reported in

Table 4. CADF test results.

	I(0)
	Stat.	Prob.
NRR	−1.958	0.025
RGDPC	−2.641	0.004
SCD	−2.610	0.000
ICT	−4.552	0.000
LNPATN	−1.303	0.096
FIE	−3.914	0.000
URB	−4.702	0.000

.

When the results of Table 3 are analyzed, it is found that all variables are stationary at level I(0).

Table 5. KRLS estimation results.

	Avg.	p > t	P10	P20	P30	P40	P50	P60	P70	P80	P90
SCD	0.126	0.067 *	−0.540	−0.295	−0.141	−0.026	0.077	0.201	0.357	0.522	0.821
RGDPC	0.001	0.917	−0.081	−0.047	−0.027	−0.015	−0.0002	0.011	0.030	0.048	0.084
ICT	−0.001	0.021 **	−0.0001	−0.0002	−0.0005	−0.001	−0.002	−0.006	−0.016	−0.032	−0.048
LNPATN	−0.099	0.000 ***	−0.337	−0.217	−0.165	−0.114	−0.078	−0.036	0.001	0.045	0.110
FIE	3.122	0.000 ***	−5.594	−3.038	−1.726	−0.371	0.597	2.192	4.336	8.321	15.649
URB	−0.005	0.082 *	−0.045	−0.029	−0.021	−0.014	−0.007	−0.0006	0.010	0.023	0.036
Model Fit Statistics
Lambda	0.211	Tolerance	0.516	Sigma	6	Eff. df	117.8	R2	0.75	Looloss	268.6

Note: * p < 0.1, ** p < 0.05, *** p < 0.01. In the study, the kernel ridge least squares (KRLS) method was applied using Stata software-17. A Gaussian kernel was preferred, and 20 bootstrap repetitions were performed to increase the reliability of the model.

shows the KRLS results, and

Table 6. BSQREG estimation results.

	P10	P20	P30	P40	P50	P60	P70	P80	P90
SCD	0.010 [0.011]	0.021 * [0.012]	0.040 * [0.020]	0.045 * [0.024]	0.050 * [0.025]	0.053 * [0.024]	0.059 * [0.026]	0.063 * [0.028]	0.068 * [0.030]
RGDPC	−0.0006 [0.001]	−0.0002 [0.002]	0.0002 [0.002]	0.0004 [0.002]	−0.002 [0.004]	0.008 [0.006]	0.011 [0.009]	0.013 [0.016]	−0.016 [0.022]
ICT	−0.002 *** [0.0006]	−0.004 *** [0.0009]	−0.005 *** [0.001]	−0.006 *** [0.001]	−0.008 ** [0.003]	−0.012 ** [0.005]	−0.012 [0.009]	−0.027 ** [0.008]	−0.054 ** [0.013]
LNPATN	−0.031 *** [0.004]	−0.038 *** [0.007]	−0.043 *** [0.008]	−0.045 *** [0.004]	−0.059 *** [0.007]	−0.064 *** [0.014]	−0.102 *** [0.020]	−0.143 *** [0.018]	−0.201 *** [0.042]
FIE	0.522 ** [0.209]	1.048 *** [0.234]	1.309 *** [0.175]	1.472 *** [0.173]	1.974 *** [0.275]	2.443 *** [0.330]	2.838 *** [0.419]	4.040 *** [0.764]	6.580 ** [2.767]
URB	−0.004 *** [0.0003]	−0.005 *** [0.0005]	−0.005 *** [0.0005]	−0.006 *** [0.0009]	−0.006 *** [0.001]	−0.007 *** [0.001]	−0.006 * [0.003]	−0.013 *** [0.005]	−0.003 [0.001]
Pseudo R2	0.08	0.09	0.11	0.11	0.11	0.11	0.11	0.11	0.11

Note: * p < 0.1, ** p < 0.05, *** p < 0.01.

shows the BSQREG results, which test the robustness of the KRLS results.

According to the KRLS analysis, the SCD variable was found to be positive and statistically significant at the 10% level of significance. While the effect of the RGDPC variable was found to be insignificant, the ICT variable was found to be negative and significant at the 5% level. The LNPATN variable was found to be negative and has a very strong significance at the 1% level. The FIE variable was found to be positive and significant at the 1% level, and the URB variable was found to be significant at the 10% level with a negative effect. The explanatory power of the model was high, and the R² value was calculated to be 0.75.

The SCD variable was positive in all quantiles and generally significant at the 10% level. The RGDPC variable, on the other hand, gave insignificant results in all quantiles. The ICT variable was negative in all quantiles and mostly significant at the 1% to 5% level. The LNPATN variable was also negative in all quantiles and significant at the 1% level. The effect of the FIE variable was positive and highly significant in all quantiles, while the URB variable was negative in most quantiles and significant at the 1% level. The pseudo-R² values varied between 0.08 and 0.11 between the P10 and P90 quantiles.

Table 7. Model comparison and consistency table.

Variable	KRLS Result	BSQREG Result	Consistency
SCD	Positive, significant at 10%	Positive, significant at 10% (from P20 to P90)	Consistent
RGDPC	Insignificant	Insignificant (across all percentiles)	Consistent
ICT	Negative, significant at 5%	Negative, significant at 1–5% (from P10 to P90)	Consistent
LNPATN	Negative, highly significant at 1%	Negative, highly significant at 1% (from P10 to P90)	Consistent
FIE	Positive, highly significant at 1%	Positive, significant at 1–5% (from P10 to P90)	Consistent
URB	Negative, significant at 10%	Negative, significant at 1–10% (from P10 to P80)	Consistent

compares the results of KRLS and BSQREG.

The SCD, ICT, LNPATN, and FIE variables are similar in both methods. The ineffectiveness of RGDPC and the negative effect of URB are also parallel in the two analyses. This shows the robustness of the model and the consistent course of the variable effects throughout the distribution.

Figure 3 provides a summary of the findings from the KRLS and BSQREG estimators.

4.2. Discussion

Although the findings are derived from OECD countries, the underlying mechanisms of digital exports and natural resource efficiency may hold broader relevance to emerging and developing economies. This improves the generalizability of the results beyond the sample context.

When the effect of SCD on NRR is analyzed, the KRLS and MMQREG results support each other. It is seen that an increase in SCD increases the NRR, and this situation is more pronounced in the upper quantiles. This finding also indicates that resource consumption increases with the development of transport networks. In addition, the development of the transport network increases the amount of CO₂ from logistics, disrupting the SDG and COP28 targets. According to Usman et al. [9], the transport sector, which is a critical element of supply chains, is responsible for approximately 30% of global energy consumption from fossil fuels, leading to excessive consumption of fossil fuels and increasing dependence on non-renewable resources [10]. Again, according to [41], SCD affects the use of natural resources in social, economic, and environmental aspects. The findings obtained in this context support the studies presented in [41,42].

When the effect of ICT goods exports on NRR is analyzed, it is observed that ICT goods exports decrease NRR. This result also supports the MMQREG results. ICT goods support optimal energy use by affecting environmental sustainability and the reliability of energy supply. On the other hand, ICT products may require less use of natural resources and less use of physical products compared to other sectors. Therefore, in terms of sustainability, the impacts of ICT products are more favorable than those of other industries. This finding is similar to [43,44].

Patents encourage more efficient resource use in industrial processes by promoting the commercialization of new technologies [45,46]. This contributes especially to the sustainable management of natural resources and reduces the risk of depletion [47]. In addition, the development of innovative solutions and green technologies through patents enables a reduction in environmental impacts and the more efficient use of resources [48]. In this context, the results of the study are in line with the literature, and it has been determined that patents reduce the use of natural resources.

Financial institutions can have a direct impact on NRs by investing in NRs or providing financial services to these sectors [49]. In particular, financial institutions operating in sectors such as energy, mining, and agriculture can cause these sectors to grow and thus increase the use of NRs [50]. However, this situation may lead to excessive resource utilization and may have negative consequences for sustainability. When the findings of the study are analyzed, it is found that FIE increases the use of NRs, and this situation increases along the quantiles. This finding is similar to [51,52,53]. The finding that the efficiency of financial institutions increases natural resource rent in OECD countries reflects a complex dynamic that supports economic growth but also challenges the sustainable management of natural resources. In addition, while this situation offers important opportunities for efficient use of resources and reduction in environmental impacts in OECD countries, it can also lead to environmental impacts, overexploitation of natural resources, and negative impacts on the quality of life of local communities.

When the effect of URB on NRR is analyzed, it is found that URB reduces NRR. With the urbanization process, the expansion of cities, and population growth, natural resources such as agricultural lands, forest areas, and water resources are under pressure and decreasing [54]. According to the study findings, the negative effect of the urbanization rate (URB) on natural resource efficiency (REF) is consistent with the “urbanization paradox” in the literature [55]. While urbanization is generally expected to improve resource use through technological diffusion and infrastructure efficiency, the increase in energy consumption and environmental pressure, especially in uncontrolled and rapid urbanization processes, can have a decreasing effect on efficiency. Therefore, the findings show that urbanization should be managed well, not only quantitatively but also qualitatively, to be compatible with sustainability goals.

This situation may lead to greater demand for critical natural resources, especially water resources and agricultural products, and make it difficult to manage these resources sustainably. This situation also negatively affects the income from natural resources. In this regard, our study is similar to the results of [56,57,58,59].

The findings of the study show that economic growth has a statistically insignificant effect on natural resource rent in OECD countries. This indicates that economic growth does not have a significant effect on natural resource utilization or that this effect is not measurable. This finding was obtained; while it contradicts [39,60,61,62], it is similar to [63].

5. Conclusions and Recommendations

Despite robust model specifications, this study is subject to limitations, including the possibility of omitted variable bias and the inability to establish strict causal relationships due to the observational nature of the data. Future research employing causal inference methods such as IV estimation or natural experiments is encouraged.

This study examines the effects of SCD and ICT product exports on NRR in European countries between 2004 and 2022. The findings of the study show that strong supply chains increase natural resource rents. It also reveals that ICT product exports reduce natural resource rents. Patents reduce NRR by making resource use more efficient, while investments made by financial institutions in resource-intensive sectors increase NRR. In addition, urban population growth reduces NRR by putting pressure on natural resources. On the other hand, economic growth in OECD countries does not have a statistically significant effect on natural resource rents.

In the interpretation of the findings, not only the transfer of regression outputs but the effects of variables were evaluated together with the literature, and causal mechanisms were discussed in an empirical context. For example, the negative effect of ICT products, lower physical production needs, data-oriented energy management, and transition to green technology applications are explained by mechanisms [44]. In contrast, the positive effect of SCD exports on NRR, increased physical logistics volume, has been associated with the expansion of the environmental footprint due to transportation-based energy demand and supply chains [20]. This shows that despite the positive economic contributions of the SCD, it can pose a risk factor in terms of environmental sustainability.

Another important finding in the model results is the reducing effect of urbanization on NRR. This situation, which can be contradictory at first glance, can be explained by the use of more efficient infrastructure of urbanization, the spread of resource-saving technologies, and the increase in conscious consumption patterns [54]. In this context, the finding is not only statistical; Environmental, Social, and Governance (ESG) dimensions have been reinterpreted in the framework. In addition, it is noteworthy that the effect of economic growth on NRR is not statistically significant. Although this finding does not mean that economic growth does not produce environmental impact, it reveals that the nature and direction of growth are decisive for environmental effects. The structural boundaries of the model were discussed in the study. Although the KRLS method provides a flexible tool for capturing non-linear effects, failure to support causality tests may limit the interpretation of the results. For this reason, advanced causality analyses and alternative modeling techniques are recommended to support future research. In addition, methodological limitations such as data constraints, country-based sample heterogeneity, and possible skipped variable prejudice are presented with transparency within the scope of the study.

Policy proposals were directly associated with the findings, avoiding general normative expressions. For example, based on the finding that ICT exports have reduced NRR, digital transformation incentives and sustainable information investments have been emphasized within the scope of policy proposals. Similarly, the finding that financial institutions’ orientation to resource-intensive sectors increases the NRR increases the importance of sustainable financing mechanisms (for example, green bonds, ESG credit systems). The finding of urbanization supports the integration of sustainable urban planning and natural resources into the policy documents of the holistic administration on the city scale. In light of these findings, the following recommendations can be made for the sustainable management of natural resources and minimizing environmental impacts:

(i) Considering the resource-efficiency-enhancing effect, R&D studies should be encouraged for the development of innovative and environmentally friendly technologies. (ii) Considering the environmental impacts of investments made by financial institutions in resource-intensive sectors, the orientation towards sustainable and low-carbon projects should be increased. (iii) To reduce the pressure of urban population growth on natural resources, sustainable water and land management strategies should be adopted in urban planning. (iv) Due to the impact of ICT products in reducing natural resource consumption, the development of this sector should be supported, and digital transformation processes should be accelerated. (v) Renewable energy use, optimization, and recycling strategies should be implemented to reduce the environmental footprint of supply chains. These recommendations will contribute to more effective and sustainable management of natural resources. Future studies are also encouraged to build on these findings by conducting similar analyses across different regions or sectors.

In addition, as a limitation of the study, although the study considered European countries collectively, there are significant regional differences in factors such as resource use, digitalization level, and GVC participation among these countries. In this study, the geography of Europe was evaluated as a homogeneous structure. Although this approach is useful in identifying general trends, it carries the risk of ignoring regional contexts. Therefore, it is recommended that future studies focus on sub-regions within Europe (e.g., Western Europe, Central and Eastern Europe, Scandinavian countries). While the KRLS and BSQREG models offer flexible estimation techniques capable of addressing non-linearity and distributional heterogeneity, they primarily identify associations rather than causal effects. Due to the observational nature of the data and the absence of valid instruments, strict causal inference was beyond the scope of this study. Future research should consider methods such as instrumental variable (IV) estimation or difference-in-differences (DID) models to better address potential endogeneity issues.