Literature Review

At the theoretical level, the relationship between economic growth and the total amount and structure of resource consumption has garnered significant attention in the field of resource economy. Since the 1990s, scholars have predominantly focused on studying the nexus between economic growth, environmental pressures, and resources, emphasizing theoretical deepening and the application of quantitative methodologies. Thus, the introduction of decoupling theories, methods, and indicators provides theoretical guidance and diversified quantitative approaches for investigating sustainable resource development. Presently, numerous studies explore resource-environment variables (e.g., CO₂, energy, urban waste) and short-term local pollutants (e.g., SO₂, SPM, NO_X, and CO), as well as land consumption within the realm of resources and the environment [1]. However, there remains a paucity of research on the decoupling conditions of non-energy mineral resource consumption.

The decoupling analysis approach endeavors to sever the connection between resource and environmental pressures and economic performance. It characterizes the intrinsic relationship between the two through quantitative relations while effectively identifying the specific phases and real-time information pertaining to their evolutionary dynamics. This approach enables an examination of the relationship between economic growth and the consumption of material resources, energy consumption, land occupation, environmental pollution, and carbon emissions. Decoupling analysis focuses on long-term and short-term real-time changes in the interaction between economic development and resource and environmental pressure [2]. Shen et al. (2017) [3] posit that China’s economic growth and mineral resource consumption exhibit periodic decoupling and recoupling. Wang et al. (2020) [4] contend that the decline in total energy consumption stems from the slowdown in economic growth since 2013, the profound adjustment of industrial structure, and the continuous decrease in energy consumption intensity. Ma et al. (2021) [5] discover that China’s industrial economic growth and energy consumption demonstrate a robust decoupling state, with energy intensity and industrial structure effects playing pivotal roles in achieving such decoupling. The “dual carbon” goal necessitates accelerated industrial structural adjustments, controlled total energy consumption, enhanced energy efficiency, and the long-term promotion of green, low-carbon, and sustainable development within the energy and mineral industry.

There are many studies on basic metal encompassing aspects of supply security (Zeng et al., 2023; Duan et al., 2021; Sun et al., 2019) [6–8], demand forecasting (Zhang et al., 2015) [9], and consumption intensity (Yang et al., 2015) [10]. Conversely, there exists relatively limited research on the influencing factors and driving forces behind consumption structure. Currently, several approaches are employed to study consumption structure and its impact factors concerning basic metal. One method revolves around input-output sensitivity analysis, using this analysis to examine the structural and technological incentives as well as the driving factors behind consumption growth based on the Leonev input-output model (Yan et al., 2015; Liu et al., 2016; Huang et al., 2018) [11–13]. This method primarily scrutinizes the interrelationships among industrial sectors, with extensive research focusing on carbon emissions. Another method examines the factors influencing long-term and short-term changes in resource consumption through per capita mineral resource consumption analysis. Malenbaum (1978), for instance, conducted a comprehensive study of the long-term shifts in demand for mineral resource consumption in the United States, based on per capita mineral resource consumption. His findings suggested an inverted “U”-shaped relationship between the intensity of mineral resource consumption and per capita income. The Institute of Mineral Resources, Chinese Academy of Geological Sciences, developed an “S” pattern model for per capita copper resource consumption, concluding that per capita GDP constitutes the primary driver for copper consumption. Further research has explored the consumption structure of copper resources in China, revealing a striking resemblance to Japan’s trend in copper consumption structure during the 1960s-1970s. Researchers employ regression models to analyze the impact of factors such as population, technology, industrial structure, and wealth on resource consumption (Wiedmann, 2015; Wang et al., 2016) [14, 15]. Econometric models offer a certain degree of explanation regarding the influence of factors like industrial structure, technology, and economic development level on resource consumption costs. However, they fail to elucidate the correlations among industrial sectors. The third approach involves factor decomposition models. This method employs index decomposition and structural decomposition techniques to disentangle the primary driving forces impacting metal consumption at the global and national levels. These forces include economic scale, industrial structure, and intensity effects (Yao et al., 2018) [16]. Factorization models are widely employed in the study of energy and carbon emissions as a primary empirical method for analyzing influencing factors.

Chen et al. (2023) [17] utilized the Kaya identity and the LMDI decomposition method to assess and analyze the implicit carbon emissions generated in the energy consumption process of 30 mainland Chinese provinces (excluding Tibet) from 2003 to 2020, considering both temporal and spatial dimensions. He et al. [18] analyzed the influencing factors of carbon emissions in China’s electricity industry based on the combination of LMDI and K-means clustering. Guo et al. [19] used the kaya identity and LMDI decomposition methods to identify the contributions of six factors to NO_X emissions reduction in China’s power industry. Quan et al. [20] used the LMDI method to analyze the influencing factors of carbon emissions in China’s logistics industry. Liu et al. [21] established LMDI and Tapio models to deeply explore the influence degree of different factors on carbon emissions. De Oliveira-De Jesus [22] and Alajmi [23] analyzed the factors of electricity-related CO₂ emissions in the Latin America and Caribbean (LAC) power sector and greenhouse gas emissions in Saudi Arabia based on the LMDI model.

Since the 1980s, Divisia decomposition has gained prominence in the fields of energy and carbon emissions due to its efficacy in addressing zero-point and residual issues during decomposition. The Divisia decomposition method circumvents the need for input-output tables while providing more rational and scientific explanatory factors, excluding unexplained residual terms. Consequently, the conclusions derived from actual research align more closely with reality and boast wider applicability. Thus, this decomposition method is extensively employed in causal analyses of energy consumption and intensity [24–26]. Fang et al. [27] and Jiang et al. [28] analyzed the influence mechanism of China’s electricity consumption and non-residential power consumption based on a multi-period ST-LMDI model and Kaya-LMDI methods. Murni Yetty et al. [29] provided econometric individual time series VEC models and discovered that economic variables were the most significant factors that aggravated the change in household energy consumption based on LMDI. Hasan and Liu [30] used the LMDI to explore the factors influencing natural gas consumption in Bangladesh from 1994 to 2018.

To address the gap in understanding decoupling between fundamental metals and the drivers of economic growth and consumption structure, this paper utilizes copper as an exemplar. To begin with, an analysis of the cumulative copper resource consumption patterns during the industrialization processes spanning over 250 years in the United Kingdom, nearly 230 years in the United States, approximately 150 years in Japan and Germany, and almost 70 years in China reveals a strikingly similar trajectory of evolutionary patterns. Moreover, by applying the Tapio Decoupling Theory, we delve into the intricate relationship between copper consumption and economic growth in major countries. Furthermore, we employ the LMDI factor decomposition model to gain deeper insights into the driving forces behind China’s copper consumption. This meticulous approach allows for an exploration of the intrinsic factors influencing the growth of mineral resources during China’s middle and late stages of industrialization. Subsequently, we meticulously decompose metal consumption along three pivotal dimensions: the scale effect, structural effect, and efficiency effect. Through this comprehensive analysis, we aim to illuminate the intensity of each effect within China’s metal consumption growth process, identify the driving factors fueling economic growth, and provide astute recommendations and countermeasures for the development of the fundamental metal industry within the framework of the “dual carbon” goal.

Tapio Decoupling Model

The Tapio decoupling elastic model possesses notable advantages, including stability, accuracy, and resilience to statistical dimensional variations [35]. In this study, we employ this model within the realm of metal and metal consumption to investigate the relationship between economic growth and metal consumption during the process of industrialization. The Tapio decoupling elastic model is constructed as follows:

Where DEI represents the decoupled elastic index, EP refers to environmental pressure indicators (in this study, it pertains to metal consumption), and DP denotes the economic driving force indicator. Specifically, we utilize per capita GDP as the economic driving force indicator, with t₀ denoting the base period and t₁ representing the current period. Given the lagged relationship between resource consumption and economic growth, a time interval of 5–10 years is necessary for period analysis. Here, we adopt a 5-year time interval. Table 1 classifies Tapio decoupling elasticity into eight levels: strong decoupling, weak decoupling, recession decoupling, strong negative decoupling, weak negative decoupling, growth negative decoupling, growth connection, and recession connection. Among these levels, strong decoupling signifies the optimal state where resource and environmental pressures gradually decrease alongside economic growth, whereas strong negative decoupling represents the worst decoupling state characterized by a negative correlation between economic growth and resource/environmental pressure. The economic implications associated with other states lie along a continuum, as resource and environmental pressures intensify concurrently with recessions.

LMDI Model

Data Collection

For this paper, the sample countries for industrial data were selected from five countries: the United Kingdom, the United States, Germany, Japan, and China. These countries represent typical developed industrial nations, with China being a developing and late industrial power. The focus of the data sample is on copper as the predominant mineral resource type. Consumption data for mineral resources in the United Kingdom, Germany, and Japan from 1848 to 1973 were obtained from the Economic Statistics of Major Capitalist Countries (1848–1960) and the Statistical Summary of the Mineral Industry published annually by the British Geological Survey. Data on mineral resource consumption in the United Kingdom, various countries, and Japan from 1974 to 2019 were sourced from the World Bureau of Metals Statistics published by the British Geological Survey in different years. Data on mineral resource consumption in the United States from 1900 to 2019 were obtained from the United States Geological Survey. Population data for the United States, United Kingdom, Germany, and Japan from 1848 to 1960 were acquired from the Economic Statistics of Major Capitalist Countries (1848–1960), while population data from 1961 to 2019 were obtained from the International Statistical Yearbook. Per capita GDP data for the United States, United Kingdom, Germany, and Japan were sourced from the Groningen Center for Economic Growth and Development (GCDC). Data on copper consumption in China were gathered from the China Nonferrous Metals Industry Association, “New China Nonferrous Metals Industry 60 Years,” and the “China Nonferrous Metals Industry Yearbook 2017.” Copper consumption data for specific industries within China were sourced from the copper market analysis report of the Antaike Research Center on China Metal Network (www.metalchina.com) over the years. It is worth noting that pre-2006 copper market analysis reports did not include industry-specific consumption data, so this study utilized copper consumption data by industry from 2006 to 2017 as the data source. Annual GDP and industry-specific GDP data were obtained from the “China Statistical Yearbook” and the official website of the National Bureau of Statistics of China.

When using the LMDI factorization method, the consumption of energy and metal across various industries is generally classified using the three-industry classification method or the “total energy consumption by industry” classification method in the annual data from the National Bureau of Statistics. Energy consumption is typically divided into agriculture, forestry, animal husbandry, fishing, industry, construction, transportation, warehousing and postal services, wholesale and retail, and other sectors. Copper, as a bulk non-ferrous metal, exhibits unique consumption patterns. The sectors consuming copper include electricity, air conditioning and refrigeration, electronics, construction, transportation, and other industries. Therefore, when applying the LMDI method for factorization, the data needs to be processed using specific approaches.

Fig. 1.

The correlation between per capita copper consumption and per capita GDP for typical countries during the process of industrialization.

Results of the Copper Consumption Decoupling Model

Industrialization serves as the main theme of a country’s modernization and constitutes the fundamental path for the ascendance of modern powers. Throughout history, industrialization has spanned over 230 years, with the United Kingdom being the world’s first industrialized nation. The United States stands as the most robust industrialized country to date, while Germany has consistently demonstrated a strong capacity for industrialization. Japan experienced significant lag until 1980 but has since witnessed rapid growth. Currently, Japan’s level of industrialization is on par with previous countries. Analyzing the evolutionary pattern of GDP growth rates, the United States reached its industrialization peak from 1950 to 1960, followed by Germany from 1960 to 1967 and Japan from 1963 to 1969. During these periods, GDP growth was the fastest, averaging 10.43% and peaking at 12.88%. After 1980, China’s GDP growth remained strong, leading to rapid economic expansion. Upon entering the middle and late stages of industrialization, China’s economy experienced a slowdown and shifted toward top-quality development.

Considering the development pattern of bulk metal alongside advancing world industrialization, it is evident that the industrialization paths of developed countries share similarities but also exhibit differences due to variations in historical conditions, political factors, economic circumstances, and national industrialization spirits. Similarly, copper consumption reflects distinct patterns influenced by factors such as resource endowments, timing of industrialization, technological conditions, and government intervention. Therefore, this study compares developed countries like the United Kingdom, the United States, Germany, Japan, and China to explore the decoupling process of copper consumption. By examining the evolution of total copper consumption and its structure, we gain insights into consumption patterns and the factors influencing copper consumption in China.

The correlation between per capita copper consumption and per capita GDP for typical countries during the process of industrialization is shown in Fig. 1. Given copper’s structural and functional properties, the electrical industry accounted for the largest proportion of copper consumption in early developed countries. Analysis of copper consumption data reveals that in developed countries, copper consumption enters a growth phase when per capita GDP reaches $3000 to $3500. During the middle stage of industrialization, metal consumption demonstrates rapid growth, with an accelerated rate of copper consumption. In developed countries, the focus of copper consumption shifts to the construction industry during the middle and later stages of industrialization. Due to copper’s structural and functional properties, the peak consumption of copper occurs in the construction sector. The peak years for copper consumption in the United Kingdom, the United States, Germany, and Japan were 1977, 1999, 1998, and 1990, respectively. An in-depth analysis of the relationship between metal consumption and the industrialization process in typical industrialized countries reveals a general trend of “low growth-rapid growth-reaching the peak-tending to decline” between per capita GDP and metal consumption.

Fig. 2.

The comparison of copper consumption intensity in typical countries during the process of industrialization.

Furthermore, an examination of consumption intensity indicators, which measure resource utilization efficiency, shows that copper consumption intensity peaked earlier in the United States, with a per capita GDP peak of $8205. Given that 48% of copper consumption in the United States is attributed to the construction sector, saturation and slowdowns in the housing market significantly impact copper consumption intensity. In Europe, Japan, and China, copper consumption accounts for a significant proportion of the electrical industry. Copper consumption intensity peaks at different per capita GDP levels for Japan ($11433), the United Kingdom ($12383), and Germany ($17299), as shown in Fig. 2. The variation in peak per capita GDP levels arises from the late entry of copper into the economic and social cycle and differences in downstream consumption structures among countries, as well as the development of alternative markets such as aluminum. The relationship between copper consumption intensity and per capita GDP in major industrial countries during the industrialization process generally exhibits a Kuznets inverted “U” curve.

Fig. 3.

Trends of the Copper Decoupling Index in major industrial countries from 1980 to 2020.

Based on Equation (1), this study calculates the decoupling index between copper consumption and economic growth in China, the United Kingdom, the United States, Germany, and Japan, and presents the trends of the decoupling index in Table 2. For the entire period from 1970 to 2020, copper consumption and economic growth in the UK and the US show a strong decoupling state, achieving complete decoupling. However, copper consumption and economic growth in Germany and Japan display periodic decoupling and recoupling phenomena.

Currently, China’s copper consumption remains in the pre-peak climbing stage (see Fig. 3). With the exception of 1985–1990, a negative decoupling relationship between copper consumption and economic growth is observed in Fig. 3. The decoupling elasticity of copper consumption exceeds 1.5 and exhibits a gradual downward trend, indicating a negative decoupling state of growth. This suggests that China’s economic growth heavily relies on metal resources, but resource utilization is improving. As copper consumption is still in the growth phase and has not yet reached its peak, it can be inferred that China’s consumption of metal minerals will continue to grow at a certain rate in the future as it strives to become a strong manufacturing and quality-focused country. The “dual carbon” goal will impose mandatory constraints on high energy-consuming industries like non-ferrous metals, altering the timeline and decoupling trajectory for China’s peak copper consumption. The utilization of technology and industrialization of existing minerals, such as “urban mines,” will also highlight technological and economic benefits. Studying the internal drivers of copper consumption enables us to better comprehend the relationship between copper consumption and economic growth, as well as the connection between changes in industrial structure and improvements in resource utilization efficiency.

Analysis of LMDI Decomposition Results

Analysis of the Total Consumption and Consumption Structure of Copper in China

As a bulk non-ferrous metal, copper plays a significant role in various industries such as electrical, light industrial, mechanical manufacturing, construction, and defense. Its consumption is closely linked to the industrialization process of a country. Over the years, China’s total consumption of copper has exhibited consistent growth since the establishment of the People’s Republic of China. From 2007 to 2010, there was a rapid growth period for copper consumption, with a peak growth rate of 25.93% in 2010. However, since 2012, the growth rate has gradually slowed down.

Fig. 4.

Decomposition results of the annual effects of copper consumption in China from 2007.

In terms of industry structure, the power sector and air conditioning and refrigeration industries are the major sub-sectors driving copper consumption in China. The power sector has consistently accounted for over 40% of copper consumption since 2006, and this proportion is projected to reach 52% by 2018 due to increased investment in power grid construction and production in industries like air conditioning. Additionally, the transportation, electronics, and construction industries contribute to copper consumption but to a lesser extent; hence, they have been merged into the “other” category. Similarly, the power, air conditioning, refrigeration, and electronics industries have been grouped under the “industrial” category.

Decomposition of Factors Influencing Changes in Copper Metal Consumption in China

To analyze the factors driving China’s copper consumption from 2007 to 2018, the LMDI model was utilized, taking 2006 as the base period. This paper calculates the annual and cumulative effects of GDP per capita, industrial structure, and resource consumption intensity. Fig. 4 depicts the cumulative effect decomposition plot and the contribution of each factor. The results indicate that, with 2006 as the base period, the cumulative effect of China’s copper consumption in 2017 was negative for structural and efficiency factors but positive for scale factors. This implies that changes in industrial structure and resource consumption intensity have somewhat hindered the growth of copper consumption, while the increase in GDP per capita has had a positive impact. In terms of contribution, the cumulative effect of GDP per capita remains the largest during the middle and later stages of industrialization, with the scale effect contributing 134.31% to the cumulative increase in copper consumption in 2017. On the other hand, the cumulative contributions of industrial structure and resource consumption intensity were negative, contributing -14.62% and 19.69%, respectively, to the cumulative increase in copper consumption.

The main driving force behind the growth of copper consumption in the middle and later stages of China’s industrialization is the increase in GDP per capita. This paper computes that the scale effect consistently demonstrates positive values in both annual and cumulative impacts, signifying a robust correlation between rising GDP per capita and increased copper consumption. The contribution of the scale effect is also significant, emphasizing the substantial influence of GDP per capita on copper consumption. This conclusion is further supported by Figs. 4 and 5. The annual scale effect factor accounted for 90.50% of the total change in copper consumption in 2022, while the cumulative scale effect factor from 2007 to 2022 contributed to 129.84% of the total change in copper consumption, making it the primary explanatory factor for the increase in copper consumption.

Fig. 5.

Decomposition results of the cumulative effects share of copper consumption in China from 2007 to 2022.

Overall, the industrial structure effect and consumption intensity effect have consistently shown negative values during this phase, indicating that these factors have primarily suppressed the growth of copper consumption. Fig. 5 presents the cumulative effect values for each factor, revealing that the efficiency effect had a greater inhibitory effect on copper consumption from 2007 to 2014, while the structural effect had a stronger impact from 2014 to 2017. The regulatory impact of the efficiency effect has significantly intensified from 2018 to 2022. The negative regulatory influence of the structural effect suggests that changes in industrial structure have impeded the increase in copper consumption in China. Conversely, the inverse regulatory effect of the efficiency effect indicates that technological innovation has led to a decrease in copper consumption. The declining proportion of industrial GDP in China from 2006 has had a certain inhibitory effect on the growth of copper consumption. This is attributed to the optimization and upgrading of the industrial structure, the shift in the proportion of the service sector relative to manufacturing, and the development of refined manufacturing, all of which have suppressed copper consumption. In terms of contribution, the annual effect contribution of the industrial structure is relatively low compared to other factors, indicating that changes in industrial structure primarily exert a negative regulatory effect on copper consumption during China’s middle and later stages of industrialization. Since 2014, this negative regulatory effect has been increasing year by year.

Robustness Analysis of the Measurement Model

The LMDI decomposition method disaggregates the determinants of copper consumption into three key factors: per capita GDP, industrial structure, and resource consumption intensity. To assess the reliability of the model, a robustness analysis of the LMDI decomposition is necessary. The primary objective of this robustness analysis is to evaluate the credibility and stability of the model’s outputs. Regression analysis, a widely accepted method, is employed to assess the impact of different independent variables on the dependent variable, allowing for a comparison between the regression results and the outputs of the LMDI decomposition.

In line with the framework of our measurement model [i.e., Eq. (2)], we applied the STIRPAT model to conduct the regression analysis [38]. This approach allowed us to determine the extent to which changes in these factors affect overall copper consumption.

Taking the natural logarithm of both sides yields,

The random-form STIRPAT model permits the incorporation of additional influencing factors to analyze the elastic relationship between the explanatory variable and the dependent variable. Factor T is further disaggregated in the model with additional control variables incorporated to examine their impact on I, and regression analysis is performed to enhance the model’s applicability. In this paper, the model is applied to analyze the influencing factors of copper metal consumption, including economic growth, industrial structure, resource consumption intensity, urbanization, and technological progress. Here, C represents the consumption of mineral resources; A is the economic growth variable, indicated by per capita GDP; INTR denotes the industrialization rate, expressed by the proportion of industrial added value to GDP; CIT signifies the urbanization rate, represented by the proportion of the non-agricultural population; INTE measures the consumption intensity of mineral resources, indicated by the amount of mineral resource input per unit of GDP; TEC reflects technological progress, represented by the proportion of high-tech product exports to manufactured product exports; and PRICE refers to the price of mineral resources.

Fig. 6.

Ridge trace of copper consumption in China.

To mitigate the impact of heteroscedasticity, the model is linearly transformed using logarithms, yielding the following model:

To ensure the effectiveness and accuracy of the model results, mitigate the influence of multicollinearity on regression outcomes, and enhance the stability and reliability of traditional ordinary least squares (OLS) estimation, ridge regression with partial estimation was employed for model fitting. Ridge regression significantly improves the mean square error of traditional least squares estimation, enhances the numerical stability of model estimation at the cost of some information loss, and yields more reliable regression coefficients than OLS. The ridge regression function was used to fit the model, and the ridge regression coefficient demonstrating gradual stability was selected according to the ridge trace plot, as shown in Fig. 6.

As evident from the ridge trace plot, when the ridge parameter (k) ranges from 0 to 1 with a step size of 0.01 and the ridge regression coefficient is set at 0.2, the regression coefficients of each explanatory factor tend to stabilize. The standard ridge regression equation is:

With China’s gradual integration into the global industrialized economy, copper consumption has increased steadily, with the rapid expansion of economic scale at this stage exhibiting a pronounced “compression” evolution trend. The standard ridge regression results indicate that economic growth is the primary driver of copper consumption growth, with an overall contribution ratio of 0.2481. The regression coefficients for the industrial structure and consumption intensity factors on copper consumption are negative, showing a certain inhibitory effect, though the impact is modest, at -0.1357 and -0.1157, respectively. According to the measurement results, the industrial structure has emerged as a structural factor reducing copper consumption, while the consumption intensity reflects the efficiency of copper resource utilization, which also somewhat restrains the growth of copper consumption.

The research conducted using the modified STIRPAT model not only validates the empirical conclusions drawn from the LMDI factor decomposition but also delves deeper into analyzing the impact levels of factors such as per capita GDP, industrial structure, and consumption intensity on copper consumption.