TY - JOUR
AU - Sharif, Kashif
AB - 1. Introduction Inflation indirectly leads to pollution in any country through increased production costs and altered consumption patterns. Government expenditures impact pollution through funding environmental regulations, green initiatives, infrastructure, and pollution cleanup efforts. While promoting economic growth, industrial development contributes to pollution through emissions, waste, land use changes, and deforestation. According to [1], gross national expenditures on R&D for technological innovation increase CO2 emissions. Inflationary pressures have implications for various sectors of the economy, including energy production and consumption. The inflation due to heavy expenses causes the excess use of fossil fuels, which results in the growth of CO2 emissions [2]. Industrial activities are often associated with significant energy consumption and CO2 emissions. By analysing the impact of industrial growth on CO2 emissions, policymakers can identify strategies to strike a balance between economic development and environmental sustainability. Manufacturing agglomeration and total factor energy efficiency have a U-shaped connection with ecological systems [3]. The study objectives include exploring potential nonlinear and asymmetric effects among these variables. Low nutrient use efficiency has been associated with environmental degradation due to uneven fertiliser application and global usage [4]. The overuse of agrochemicals in agriculture degrades natural resources and threatens ecosystems [5]. The study aims to understand how inflation, national government expenditures, and industrial development increase or decrease environmental pollution in China, a fast-growing world economy. The research entails evaluating causation, recognising policy implications, and analysing data to guide sustainable behaviours. It’s an interdisciplinary strategy that considers geographical and temporal differences while supporting sustainable development and lowering pollution. The study examined the effects of inflation, national government expenditures, and industrial development on China’s carbon emissions from 1960 to 2022. The NARDL model is also used to account for projected asymmetries in stock market reactions to shifting inflation, national spending, and industrial development effects on environmental degradation. The study should distinguish between the effects of inflation and industrial growth on various forms of environmental pollution. Examining how inflation and economic development affect particular industrial sectors is crucial. It is important to consider how societal attitudes, consumer behaviour, and corporate accountability influence how inflation and industrial expansion affect environmental degradation. The study is exclusive in that it is the first to employ time-series analyses with long-range time-series data in a single country (China) from 1960 to 2022 to apply the ARDL and NARDL models. The results reveal a dynamic relationship between GNE, inflation, industrial growth, and CO2 emissions. Inflation increases China’s energy consumption, industrial growth, and transportation-related environmental pollution. The results show that government national expenditure affects CO2 emissions, and the GNE mitigates climate change and reduces greenhouse gas emissions. The findings also show that China’s CO2 emissions have significantly grown due to industrial expansion, which greatly influences CO2 emissions. Agricultural fertiliser consumption harms CO2 emissions, so sustainable agricultural practices are essential. Promoting efficient nutrient management and reducing nutrient runoff also helps minimize fertiliser consumption’s impact on environmental pollution. The study suggests that the industrial development of globalization and increased urbanisation have contributed to rising CO2 emissions in China. The subsequent portion of the investigation comprises four distinct phases. Following this, Section 2 provides an overview of the literature study. Section 3 delineates the research technique employed in the study, whereas Section 4 expounds on the outcomes and subsequent comments derived from the data analysis. Section 5 provides an overview of the results drawn from the study and offers ideas for further research and improvements. 2. Literature review The US and German governments’ national spending on encouraging R&D initiatives targets lowering CO2 emissions [6]. An increase in government national expenditures significantly increases forest land clearing for agricultural production in the short run, leading to more CO2 emissions [7]. The amount of money a country spends on R&D is correlated with how much carbon dioxide it produces [8]. Government national expenditures on research and development to increase industrial growth and industrial development also increase environmental pollution [9]. According to [10], gross national expenditures on R&D investments may only sometimes lead to higher CO2 emissions. [11] determined that the disparities in government national expenditures were a significant factor in national CO2 emissions. Government national expenditures on renewable energy and exports have reduced CO2 emissions [12]. The inflation rate and building material prices decline while CO2 emissions grow [13]. According to [14], inflation considerably reduces GDP growth, which causes an increase in environmental pollution. According to [15], while more inflation is good for the government’s gross domestic product, it also leads to more pollution as businesses and farms switch to cheaper fuel sources. There is no substantial correlation between inflation and renewable energy, but over the long run, renewable energy harms inflation in the short run [16]. The decrease in inflation rates has been increasing renewable energy sources, and the increase in renewable energy sources also decreases CO2 emissions in the long run [17]. According to [18], the increase in the cost of renewable energy sources also decreases CO2 emissions in the short and long run. According to [19], the increase in R&D expenditures for technological innovations also upset the CO2 emissions in China. The use of fertilisers in agriculture production increases the quantity of NO2 in soil and air, which is also a major source of environmental pollution [20]. Higher fertiliser consumption has boosted Nepal’s short- and long-term carbon dioxide emission levels [21]. Using fertilisers in the Guangdong region of China has increased environmental pollution [22]. Fertiliser consumption in crop production is very important in increasing CO2 emissions [23]. The increase in fertiliser consumption increases CO2 emissions, and the increase in renewable energy sources due to urbanisation and industrial growth lowers CO2 emissions in developed countries [24]. The continuous use of agricultural fertilisers in crop production creates soil degradation and air pollution [25]. Financial assistance for agriculture will significantly impact the use of chemical fertilisers and carbon emissions in the agricultural sector [26]. Overusing chemical fertilisers causes increased CO2 emissions from agricultural output [27]. Fertiliser use and agricultural employment decreased CO2 emissions in the long run [23]. Fertiliser consumption and livestock production significantly increased CO2 emissions in the short and long run [28]. According to [29], ARDL econometrics support the EKC hypothesis and the long-run relationship between industrial expansion and CO2 emissions. Modernising industry development increases CO2 emissions due to higher energy demand and supply chain changes [30]. According to industrial growth, CO2 emissions have a U-shaped connection in the short- and long-run [31]. Upgrading industrial structures reduces CO2 emissions [32]. Energy consumption, urbanisation, and economic expansion boost Pakistan’s CO2 emissions from industrial development [33]. Industrial investment in China is the biggest cause of rising CO2 emissions [34]. Industrial expansion and fossil fuel use are major contributors to regional CO2 emissions [35]. Renewable energy, bootstrap autoregressive distributed lag testing (ARDL), and nonlinear ARDL enhance environmental quality in the short and long term [36]. The ARDL and NARDL symmetric analyses show that economic growth increases CO2 emissions, but crude oil prices and FDI inflows increase CO2 emissions [37]. The baseline ARDL and NARDL techniques used in the research revealed that economic growth impacts CO2 emissions [38]. According to [39], the ARDL and NARDL models determined that a rise in economic growth would reduce CO2 emissions, while a decrease in economic growth would raise CO2 emissions. [40] has applied the NARDL model and determined that population density and GDP per capita increase carbon emissions in the short and long run, while income inequality does not impact carbon emissions in the short run. In the ARDL model, economic growth increases energy consumption, and urbanisation increases CO2 emissions [41]. 3. Research methodology 3.1 Data collection The study has applied data from 1960 to 2022, which has been collected from the World Bank database "https://databank.worldbank.org.com” URL https://databank.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG/1ff4a498/Popular-Indicators” for CO2 emissions, inflation (IN), and fertiliser consumption (FR). The data for gross national expenditure (GNE) and industrial growth (IG) has been collected from the International Monetary Fund Database ("https://www.imf.org.com)," URL “https://climatedata.imf.org/pages/access-data” the World Economic Outlook and China Premium Database ("https://www.ceicdata.com)" URL “https://info.ceicdata.com/en/en-products-global-database.”. China is a very fast-growing economy in the world, and due to its high industrial and agricultural production levels, it is the largest producer of CO2 emissions. China’s CO2 emissions have a significant impact on global climate change. Understanding the factors contributing to these emissions is critical for addressing environmental concerns. China’s economy is closely linked to its industrial growth and gross national expenditures. Understanding the relationship between these factors and CO2 emissions can provide insights into the potential economic impacts of efforts to reduce emissions. Table 1 demonstrates the variables’ symbols, units, and descriptions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 1. Displays the variables’ detail. https://doi.org/10.1371/journal.pone.0297413.t001 Understanding the global economic and environmental difficulties requires examining the connection between China’s Gross National Expenditures, inflation, industrial expansion, and CO2 emissions. 3.2 Research design Designing a clean and sustainable energy system is a complex task that requires a multidisciplinary approach, and the investigation uses the theoretical links to come up with the equation: (1) The study employed CO2 emissions as the dependent variable, whereas the independent variables are inflation, gross national product (GNE), industrial development, and fertilizer consumption. The research anticipates the following connections between these variables: More pronounced degrees of ecological innovation correlate with lower fossil fuel by-products. Greater GNE results in higher CO2 emissions. The story of CO2 emissions increases when FP levels rise. IG use increases are correlated with decreased CO2 emissions. The smaller the CO2 emissions, the higher the FR. 3.3 Econometrical background The NARDL model’s graphical structure looks like this: Using standardized unit root tests, such as the Augmented Dickey-Fuller (ADF) test, establishes the order of integration of the variables. The generated graphical framework illustrates the causal connections between the NARDL model’s variables. Fig 1 demonstrates the graphical framework of the nonlinear short ARDL-bounds testing approach (NARDL). Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 1. Illustrates the graphical framework of the research methodology. https://doi.org/10.1371/journal.pone.0297413.g001 The NARDL approach allows both short-run and long-run dynamics in the data and allows nonlinear shorts to be captured in the data-generating process. To use the NARDL method, you have to estimate an autoregressive distributed lag (ARDL) model with lagged values for the dependent and independent variables and lagged differences between the variables. According to [42], a short- and long-run nexus between industrial growth and CO2 emissions was identified using the ARDL bounds testing model. The significant difference between the traditional ARDL model and the NARDL model is the inclusion of a lagged squared term of the dependent variable in the model. Phase 1: Unit root test The Augmented Dickey-Fuller (ADF) test has been used to examine the impact of the unit root of gross national expenditures (GNE), inflation, fertiliser, and industrial growth on CO2 emissions. ADF tests the null hypothesis that a time series has a unit root to determine its non-stationarity. The null hypothesis is rejected, and the time series is considered stationary in Eq 2: (2) The series is viewed as fixed if the coefficient of the above condition approaches one, as shown by the AR (1) process At−1 is less than 1. If |φ| < 1, the data is stationary at level 1, and if |φ| = 1, the variable is non-stationary at the level and includes the unit root process. In the third scenario, |φ| > 0 for the series never achieves equilibrium because it is diverse and explosive. Phase 2: ARDL model The ARDL (Autoregressive Distributed Lag) model is a popular econometric approach used to analyze the long-term relationships among variables. The study has applied ARDL model to analyze the influence of gross national expenditures, inflation, fertilizer consumption, and industrial growth on CO2 emissions. The ARDL model regression has represented as in Eq 3: (3) Factor cointegration is optional, and F-statistics evaluate the hypothesis. Suppose the F-test result is greater than the factors cointegration. SBC and AIC will resolve slack time, and the long-run Eq 4 is as follows: (4) The short-run equation will be approximated if a long-run connection between these variables exists. The error correction run will be negative and demonstrate convergence to equilibrium. (5) Phase 3: NARDL model The study used the [43] NARDL model to detuning how each response variable responds to both positive and negative shocks and evaluate variable asymmetry in the short- and long-run. NARDL specifies the following equation in short- and long-run asymmetries: (6) Where α0, α1 … α9 are a set of estimated long-run parameters and based on above the preparation mentioned by ‘γ’ lessening in CO2 emissions. The supplementary coefficients that designate the particular connection among the NARDL model and other variables as: (7) The following describes the ECM examines the consistency of long-run parameters and the short- run consequences: (8) The NARDL model follows the stages and γ1 = γ+ = γ− = 0, presents a different theory proposes that cointegration exists in long-run. If it falls in the range between the upper and lower boundaries represent the difficulty in making a choice and is referred to as another cointegration method. The following hypothesis is being tested by the Wald test, i.e., or . Phase 4: The granger-toda-yamamoto principle Toda-Yamamoto model dynamic VAR(p) model is written as: (9) Granger causality from GNE to CO2 implies that P11j = 0, and Granger basis from CO2 to GNE implies that P66j = 0. After looking at both short-run and long-run parameter estimates, the study used the Granger causality test to determine the hypotheses. Vector auto-regression (VAR) and vector error-correction models (VECM) were used to find the direction of Granger causality between GNE and CO2 emissions over time. These models have been used to analyze the dynamic relationship between all the variables over time and test for the presence of Granger causality. Phase 5: Innovation accounting matrix According to [44], IRF measures the dynamic interaction between the relevant variables across time and converts the VAR model. The primary application of VDA is to clarify the relative importance of variables. The Innovation Accounting Matrix has been applied to track and measure progress in innovation related to the influence of gross national expenditures, Inflation, Fertilizer Consumption, and Industrial Growth on CO2 Emissions. Gross National Expenditures, Inflation, Fertilizer Consumption, and Industrial Growth are typically used as economic indicators to measure a country’s economic performance. CO2 emissions measure the amount of carbon dioxide emissions into the atmosphere. 3.1 Data collection The study has applied data from 1960 to 2022, which has been collected from the World Bank database "https://databank.worldbank.org.com” URL https://databank.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG/1ff4a498/Popular-Indicators” for CO2 emissions, inflation (IN), and fertiliser consumption (FR). The data for gross national expenditure (GNE) and industrial growth (IG) has been collected from the International Monetary Fund Database ("https://www.imf.org.com)," URL “https://climatedata.imf.org/pages/access-data” the World Economic Outlook and China Premium Database ("https://www.ceicdata.com)" URL “https://info.ceicdata.com/en/en-products-global-database.”. China is a very fast-growing economy in the world, and due to its high industrial and agricultural production levels, it is the largest producer of CO2 emissions. China’s CO2 emissions have a significant impact on global climate change. Understanding the factors contributing to these emissions is critical for addressing environmental concerns. China’s economy is closely linked to its industrial growth and gross national expenditures. Understanding the relationship between these factors and CO2 emissions can provide insights into the potential economic impacts of efforts to reduce emissions. Table 1 demonstrates the variables’ symbols, units, and descriptions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 1. Displays the variables’ detail. https://doi.org/10.1371/journal.pone.0297413.t001 Understanding the global economic and environmental difficulties requires examining the connection between China’s Gross National Expenditures, inflation, industrial expansion, and CO2 emissions. 3.2 Research design Designing a clean and sustainable energy system is a complex task that requires a multidisciplinary approach, and the investigation uses the theoretical links to come up with the equation: (1) The study employed CO2 emissions as the dependent variable, whereas the independent variables are inflation, gross national product (GNE), industrial development, and fertilizer consumption. The research anticipates the following connections between these variables: More pronounced degrees of ecological innovation correlate with lower fossil fuel by-products. Greater GNE results in higher CO2 emissions. The story of CO2 emissions increases when FP levels rise. IG use increases are correlated with decreased CO2 emissions. The smaller the CO2 emissions, the higher the FR. 3.3 Econometrical background The NARDL model’s graphical structure looks like this: Using standardized unit root tests, such as the Augmented Dickey-Fuller (ADF) test, establishes the order of integration of the variables. The generated graphical framework illustrates the causal connections between the NARDL model’s variables. Fig 1 demonstrates the graphical framework of the nonlinear short ARDL-bounds testing approach (NARDL). Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 1. Illustrates the graphical framework of the research methodology. https://doi.org/10.1371/journal.pone.0297413.g001 The NARDL approach allows both short-run and long-run dynamics in the data and allows nonlinear shorts to be captured in the data-generating process. To use the NARDL method, you have to estimate an autoregressive distributed lag (ARDL) model with lagged values for the dependent and independent variables and lagged differences between the variables. According to [42], a short- and long-run nexus between industrial growth and CO2 emissions was identified using the ARDL bounds testing model. The significant difference between the traditional ARDL model and the NARDL model is the inclusion of a lagged squared term of the dependent variable in the model. Phase 1: Unit root test The Augmented Dickey-Fuller (ADF) test has been used to examine the impact of the unit root of gross national expenditures (GNE), inflation, fertiliser, and industrial growth on CO2 emissions. ADF tests the null hypothesis that a time series has a unit root to determine its non-stationarity. The null hypothesis is rejected, and the time series is considered stationary in Eq 2: (2) The series is viewed as fixed if the coefficient of the above condition approaches one, as shown by the AR (1) process At−1 is less than 1. If |φ| < 1, the data is stationary at level 1, and if |φ| = 1, the variable is non-stationary at the level and includes the unit root process. In the third scenario, |φ| > 0 for the series never achieves equilibrium because it is diverse and explosive. Phase 2: ARDL model The ARDL (Autoregressive Distributed Lag) model is a popular econometric approach used to analyze the long-term relationships among variables. The study has applied ARDL model to analyze the influence of gross national expenditures, inflation, fertilizer consumption, and industrial growth on CO2 emissions. The ARDL model regression has represented as in Eq 3: (3) Factor cointegration is optional, and F-statistics evaluate the hypothesis. Suppose the F-test result is greater than the factors cointegration. SBC and AIC will resolve slack time, and the long-run Eq 4 is as follows: (4) The short-run equation will be approximated if a long-run connection between these variables exists. The error correction run will be negative and demonstrate convergence to equilibrium. (5) Phase 3: NARDL model The study used the [43] NARDL model to detuning how each response variable responds to both positive and negative shocks and evaluate variable asymmetry in the short- and long-run. NARDL specifies the following equation in short- and long-run asymmetries: (6) Where α0, α1 … α9 are a set of estimated long-run parameters and based on above the preparation mentioned by ‘γ’ lessening in CO2 emissions. The supplementary coefficients that designate the particular connection among the NARDL model and other variables as: (7) The following describes the ECM examines the consistency of long-run parameters and the short- run consequences: (8) The NARDL model follows the stages and γ1 = γ+ = γ− = 0, presents a different theory proposes that cointegration exists in long-run. If it falls in the range between the upper and lower boundaries represent the difficulty in making a choice and is referred to as another cointegration method. The following hypothesis is being tested by the Wald test, i.e., or . Phase 4: The granger-toda-yamamoto principle Toda-Yamamoto model dynamic VAR(p) model is written as: (9) Granger causality from GNE to CO2 implies that P11j = 0, and Granger basis from CO2 to GNE implies that P66j = 0. After looking at both short-run and long-run parameter estimates, the study used the Granger causality test to determine the hypotheses. Vector auto-regression (VAR) and vector error-correction models (VECM) were used to find the direction of Granger causality between GNE and CO2 emissions over time. These models have been used to analyze the dynamic relationship between all the variables over time and test for the presence of Granger causality. Phase 5: Innovation accounting matrix According to [44], IRF measures the dynamic interaction between the relevant variables across time and converts the VAR model. The primary application of VDA is to clarify the relative importance of variables. The Innovation Accounting Matrix has been applied to track and measure progress in innovation related to the influence of gross national expenditures, Inflation, Fertilizer Consumption, and Industrial Growth on CO2 Emissions. Gross National Expenditures, Inflation, Fertilizer Consumption, and Industrial Growth are typically used as economic indicators to measure a country’s economic performance. CO2 emissions measure the amount of carbon dioxide emissions into the atmosphere. 4. Results and discussions 4.1 Descriptive statistics Table 2 demonstrates the descriptive statistics for CO2 emissions, fertilizer consumption (FP) (percentage of output), and fertilizer consumption (FR) (kg per hectare of arable land). Descriptive statistics show that inflation (IN) and gross national expenditures (GNE) achieved their highest levels. The distribution of industrial growth (IG) (annual growth) as a percentage of overall GDP growth is adversely skewed. The study has a sample of data on CO2 emissions, fertilizer consumption (FP), and fertilizer consumption (FR). Download: PPT PowerPoint slide PNG larger image TIFF original image Table 2. Descriptive statistics. https://doi.org/10.1371/journal.pone.0297413.t002 While overall GDP is growing, the industrial sector is growing at a different pace or may even be contracting. The interpretation of the distribution’s means, median and skewness depends on the specific context and the goals of the analysis. In some cases, an adversely skewed distribution may be desirable, such as when trying to reduce economic inequality and more sustainable economic growth. A more symmetrical distribution is preferable, aiming for balanced economic growth across all sectors. Table 3 demonstrates the 1st difference level of IG, FP, and GNE non-stationary. In time series analysis, the factors’ stationarity is essential because the non-stationarity of a unit root in the factors prompts erroneous relapse examinations. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 3. Demonstrates the unit root test results. https://doi.org/10.1371/journal.pone.0297413.t003 The results demonstrate that the data contain a unit root; the IG is non-stationary and may show a trend or drift over time. FP and IN exhibit non-stationary unit roots, indicating temporal drift. GNE data has a trending non-stationary unit root, and the non-stationary ADF test shows a unit root for CO2 emissions. Rising oil prices cut CO2 emissions, limit high-carbon energy use, increase renewable power demand, and lower real GDP [45]. The investigation may proceed to the ARDL and NARDL testing procedures given the ARDL test conditions, no second request coordinated, or I (2) variable is found. An acceptable VAR lag period is necessary before employing the ARDL bound testing approach. Table 4 demonstrates an acceptable lag length for the VAR Lag Order Selection Criteria. Vector Auto-regression (VAR) models have been used to model the dynamic relationship between the study time series variables. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 4. Demonstrates the VAR lag order selection criteria. https://doi.org/10.1371/journal.pone.0297413.t004 The AIC is a measure of the relative quality of a statistical model for the set of data. The BIC is similar to the AIC but places a higher penalty on models with other parameters. The HQIC is a modification of the AIC that provides a more accurate estimate of the optimal lag order for VAR models. The FPE is another measure of the quality of a statistical model that penalizes models with other parameters. 4.2 Wald test parameters Table 5 demonstrates the Wald test results, a statistical test used to determine whether a group of parameters in a regression model are jointly significant. In the context of a VAR model, the Wald test has been used to test the joint significance of the coefficients of multiple lagged values of different time series variables, including Gross National Expenditures (GNE), inflation (IN), fertilizer consumption (FP), and industrial growth (IG), on CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 5. Demonstration of wald test parametric statistical analysis. https://doi.org/10.1371/journal.pone.0297413.t005 The results demonstrated the influence of GNE, FR, IG, IN, and IG on CO2 emissions. The Wald test has been used for the joint significance of coefficients in the ARDL and NARDL models. To perform the Wald test, the study first estimates the ARDL model. The study used the Wald test statistic to calculate the p-value for the test, and the p-value is less than a chosen significance level of 5%. Where C is a vector of the estimated coefficients of the interaction terms, and R is a matrix of the estimated covariance of the coefficients of the interaction terms. According to [46], government expenditures on R&D for renewal energy cause a decrease in CO2 emissions, but due to high expenditures, Chinese companies are still using fossil fuels, which increases the CO2 emissions. The Wald test further demonstrated that ARDL and NARDL models depict relationships between data properties. 4.3 ARDL model and bounds test Table 6 demonstrates the [47] ARDL model in various orders, I (0) and I (1), in the short run. The ARDL model can handle data that are stationary in any order (I(0) or I(1)). The study has applied the [47] Autoregressive Distributed Lag (ARDL) model of order (2, 0, 3, 1, 1, 1) to estimate the short-run and long-run effects of independent variables (FP, FR, IG, IN and GNE) on a dependent variable (CO2 emissions). Download: PPT PowerPoint slide PNG larger image TIFF original image Table 6. Displays the ARDL model estimations for all variables in the short run. https://doi.org/10.1371/journal.pone.0297413.t006 The study has analyzed the coefficients of FP, FR, IG, IN, and GNE influence CO2 emissions. The results look at the sign and magnitude of each coefficient of independent variables FP, IG, IN and GNE statistical significance (using t-tests or p-values) to see whether these variables positively affect CO2 emissions. The mechanism’s effectiveness in promoting sustainable corporate development; however, implementing the mainland-HK Stock Connect has primarily boosted the leading firms [48]. Table 7 demonstrates the ARDL model’s long-run assessments with the bound test. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 7. Demonstration of ARDL model and Bounds Test results in the long run. https://doi.org/10.1371/journal.pone.0297413.t007 The long-run coefficients provide information on the long-term impact of changes in the FP, FR, IG, IN, and GNE influences on CO2 emissions. The long-run assessments with the bound test involve examining the long-run coefficients and statistical significance. A statistically significant coefficient reveals the long-term influence of FP, FR, IG, IN, and GNE on CO2 emissions. The government expenditures on energy technologies that meet the electricity demand, considering economic and environmental parameters [49]. The environment’s quality is influenced over the long run by IG, GNE, FR, and FP impacts on CO2 emissions are significantly and negatively impacted by IN, GNE, and FP, although IN and FP directly correlate to CO2 emissions. The industrial and resource curse concept is supported by a country’s higher GNE and IG and increased loss of natural resources, and China is eventually facing environmental pollution. Most nitrogen is lost to the environment, particularly soil, water, and air, leading to non-point source pollution [50]. Table 8 exhibits ARDL model error correction regression. Error correction evaluates the deviation from the long-run equilibrium relationship between GNE, IG, FR, IN, FP, and CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 8. ARDL model error correction regression. https://doi.org/10.1371/journal.pone.0297413.t008 The results demonstrate that the GNE, IG, FR, IN, and FP are in balance in the long run, but these are out of equilibrium in the short run. The short-run GNE is out of equilibrium, and the long-run and short-run FR & FP discrepancies are resolved within a year as IG lags in adjusting to the independent variables. According to [51], renewable energy positively impacts inflation and lowers CO2 emissions. The CO2 emissions are considerably detrained by the lag between the 1st and 3rd difference in GNE, IG, FR, IN, FP and CO2 emissions, which are significantly correlated. Even if it changed sign in the second and third lags, the short-run demand situation on IG and IN is strong and good. [52] found that moving government funding to public asset investments increases environmental pollution. The short-run impact of the stringency variable is unfavorable and negligible at the 5% significance level. 4.4 ARDL and NARDL models and F-bound test Table 9 asserts the presence of a long-run cointegration relationship by validating the bound test estimations for both ARDL and NARDL. The F-Bounds test has been applied to confirm a cointegration relationship of independent and dependent variables in the long run. The F-Bounds test is a statistical test used to determine whether cointegration exists between all the variables in an ARDL model. The F-statistic is then compared to the distribution’s upper and lower critical values, which are based on the number of variables in the model and the sample size. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 9. The f-Bounds Test confirmed a cointegration connection between the variables. https://doi.org/10.1371/journal.pone.0297413.t009 Table 10 demonstrates the NARDL model results for all the variables in the short run. The study has applied the NARDL model of order (2, 0, 0, 3, 0, 0, 0, 2, 2, 1, 0) to determine the results of all variables. It is helpful to analyze the short-run consequences of changes in emissions of GNE, IG, FR, IN, and FP using the NARDL model [53]. In the short run, the NARDL model can estimate a shock’s immediate and lagged effects on GNE, IG, FR, IN, FP and CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 10. Dynamic estimation of the NARDL Model in the short-run. https://doi.org/10.1371/journal.pone.0297413.t010 The results of the NARDL model in the short run will depend on the specific variables and data used in the analysis. The results show that the GNE, IG, FR, IN, and FP effects positively impact CO2 emissions. Inflation (IN), which is large and positive, has a considerable short-run impact on rising CO2 emissions. Between 0.03% and 0.97% of China’s GDP is used to offset the cost of inflation, with foreigners and investors bearing most of the burden [54]. Although China has agreed on the need to reduce CO2 emissions completely, there are still disparities in regional emissions. Reduced carbon emissions serve the public benefit and reveal a significant positive externality that is challenging to address in the market [11]. Due to China’s and India’s industrial revolutions, industrial growth (IG) is the primary factor driving the country’s rise in CO2 emissions in the next decades [55]. FP and FR impact short-term CO2 emissions and positively correlate with other parameters. Spreading information to fertilizer wholesalers, crop advisors, farmers, and agricultural and environmental authorities should boost BMP consumption [56]. Positive shocks in FP have a positive and substantial coefficient that greatly influences CO2 emissions. An increase in FP causes CO2 emissions to rise, whereas negative shocks to IN usage substantially impact CO2 emissions. A decrease in the usage of renewable energy sources leads to a rise in CO2 emissions, according to the negative IN coefficient. Table 11 reveals that the NARDL model long-run coefficient will converge to equilibrium. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 11. NARDL model long-run and limited constant with no trends. https://doi.org/10.1371/journal.pone.0297413.t011 The results demonstrate that the positive shocks to GNE and IG meaningfully affect fossil fuel byproducts’ high likelihood worth and negative coefficient esteem. China’s CO2 emanations are increasing over the long haul because the nation is making an ever-increasing number of farming items [57]. A positive change in IG affects CO2 discharges, showing that contamination in the climate decreases as IG improves. It is another significant component affecting fossil fuel byproducts when a nation’s funds deteriorate; because less cash is in the in the evolving hands, and pollution levels go up. Conversely, negative shocks to IG show a reasonable connection with fossil fuel byproducts. 4.5 NARDL bounds test Table 12 demonstrates the F-Bounds test to assert the cointegration connection in the short- and long-run. The NARDL model with the Bounds test has been used in econometric analysis to determine the relationship between study macroeconomic variables and CO2 emissions. Once the NARDL model is estimated, the study will conduct the F-Bounds test to assess the existence of cointegration between the variables. The F-Bounds test involves calculating the F-statistics upper and lower bounds based on macroeconomic variables’ impact on CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 12. F-Bounds test to assert the cointegration connections. https://doi.org/10.1371/journal.pone.0297413.t012 Industrial revolutions have diverse implications for achieving net-zero carbon emissions [58]. The study assessed the cointegration connection between the variables and gained insights into their long-run relationships by calculating the upper and lower bounds of the F-statistic. China’s industrial sector investment volume increases CO2 emissions [34]. 4.6 CUSUM and CUSUM of square graphs Fig 2 demonstrates the CUSUM and CUSUMSQ of the square model’s stability. The CUSUM test involves calculating a sequence of test statistics that represent the cumulative sum of the residuals from the regression model, and the results of test statistics have been plotted against the sample size and a critical value. The CUSUMSQ test involves calculating a sequence of test statistics representing the cumulative sum of the squared residuals from the regression model. The parameter stability is evaluated using the cumulative sum of recursive residuals (CUSUM) and cumulative sum of squares (CUSUMSQ) tests [59]. The study used the CUSUM and CUSUM of Squares (CUSUMSQ) tests to discover the NARDL model residuals and the link between GNE, IN, IG, FP, and CO2 emissions. Study test data were plotted against sample size and a critical value. CUSUM and CUSUM of Square tests identify time series mean shifts. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 2. The CUSUM and CUSUM of squares test results of macroeconomic variables. Source: Author’s illustration. https://doi.org/10.1371/journal.pone.0297413.g002 The graphs indicate a significant change in the mean of FP, GNE, IN, and IG related to CO2 emissions, and it suggests a potential causal relationship between FP, GNE, IN, IG and CO2 emissions. Table 13 demonstrates the diagnostic inspection test for verification of the ARDL and NARDL model results. The ARDL and NARDL models have assumed that the errors are normally distributed, and deviations from normality can lead to biased coefficient estimates [60]. The heteroskedasticity-hearty Breusch-Agnostic test is more reliable than the "wild bootstrap form" of the standardized unique Breusch-Agnostic test [61]. The diagnostic tests have been applied to identify potential problems with the ARDL and NARDL models and improve the overall performance of all the variables. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 13. Diagnostic test inspection. https://doi.org/10.1371/journal.pone.0297413.t013 The Jarque-Bera test’s probability values are more than 5%, which shows that the null hypothesis of normality is not disproved. The evidence for a U-shaped connection between the rate of industrial expansion and CO2 emission was supported by both short-run and long-run regression parameters [62]. The hidden ARDL and NARDL models’ relapses match very well; both are critical at the 1% level worldwide. These models passed the Lagrange Multipliers (LM) test for heteroskedasticity, the Jarque-Bera test, the Breusch-Agnostic Godfrey test, and the RESET test. The consumption of fertilizers and fossil fuels consumption increased environmental pollution [63]. 4.7 Estimation of a ramsey RESET model Table 14 demonstrates the RESET test confirming the results of the ARDL and NARDL models. The Ramsey RESET test is a diagnostic test applied to determine a nonlinear regression between the independent and dependent variables. The Ramsey RESET test is also a diagnostic tool used with other diagnostic tests to ensure the model is correctly specified. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 14. Ramsey RESET test with F-test summary. https://doi.org/10.1371/journal.pone.0297413.t014 It estimates the regression model of CO2 emissions as a function of Gross National Expenditures (GNE), Inflation and Industrial Growth. The results of the Ramsey RESET test have been interpreted in conjunction with other diagnostic tests and considerations, and the residuals for patterns and testing for multi-collinearity among the regressors. 4.8 Granger causality test Table 15 demonstrates the Granger causality scales between the components after the cointegration coefficients. The Granger causality test is a statistical technique used to determine whether one variable is used to predict another variable. The study has applied the Granger causality test to analyze the time series data for stationarity. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 15. Demonstrates that the Granger causality test estimations. https://doi.org/10.1371/journal.pone.0297413.t015 The results reveal that the positive shocks of GNE to CO2 emissions are correlated in both directions; negative shocks in IG, FP, and GNE Granger cause CO2 emissions. A unidirectional association between IG usage and FR confirms China’s CO2 emissions. CO2 emissions have positive shocks in IG, negative shocks in IN, and negative shocks in FP, which are all brought on by Granger. To validate the GNE theory, there is also a unidirectional association between FR, IN, GNE, and FP. Table 16 demonstrates the Granger causality test result for the effectiveness of all variables on CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 16. Granger causality test result for the effectiveness of CO2 emissions. https://doi.org/10.1371/journal.pone.0297413.t016 The results of the Granger causality test for the association between GNE, IG, and FR are intriguing. These results show that neither the Granger causality estimation from GNE to IN nor IN to IG is statistically significant. The NARDL model reveals that changes in energy usage, fertilizer use, and agricultural carbon emission lead to changes in cereal food production, both positively and negatively (Koondhar et al., 2021) [64]. The findings demonstrate that GNE, FP, IG, and FR growth will only increase CO2 emissions. Table 17 shows the impulse response function of CO2 emissions to confirm the relationship validity between the study variables. The impulse response function indicates that economic activity raises CO2 emissions mathematically. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 17. Demonstrates the Impulse Response Function of CO2 emissions. https://doi.org/10.1371/journal.pone.0297413.t017 The results demonstrate that CO2 emissions have an inverse connection with negative shocks to GNE and IG. The positive shocks to IN decrease CO2 emissions, whereas an increase in IG and GNE also increases the CO2 emissions. China’s CO2 emissions rise in response to positive shocks to FP and FR. Within a decade, green finance will significantly reduce fertilizer use and agricultural carbon emissions [65]. CO2 emissions will fall along with an increase in FP, and China’s GNE and IG will also significantly influence CO2 emissions. 4.9 Variance decomposition analysis Variance decomposition analysis has been applied to decompose the variance of the macroeconomic research variables and CO2 emissions. A quarter of all human CO2 emissions are brought on by land usage and agricultural output [66]. There are positive shocks of IG, FR, and GNE to CO2 emissions, but FP has a negative shock FP to CO2 emissions. Table 18 demonstrates the variance decomposition analysis (VDA) of CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 18. Variance decomposition analysis of CO2 emissions. https://doi.org/10.1371/journal.pone.0297413.t018 The results demonstrate that FP, GNE, IG and IN significantly impact CO2 emissions. The GNE is the most important variable, and policymakers might focus on promoting economic growth through sustainable and low-carbon technologies while reducing energy consumption and improving energy efficiency. 4.10 Discussions The findings show that greater gross national expenditures (GNE) correspond to greater economic activity, which increases CO2 emissions. Currently, no regulatory interactions link increasing economic development and reducing carbon dioxide emissions at the national level [67]. Most economic activity depends on energy usage, normally produced using fossil fuels that emit CO2. Higher inflation rates cause lower CO2 emissions. Using and manufacturing fertilizers have expanded, increasing greenhouse gas emissions. In addition, excessive fertilizer usage can result in soil deterioration and nutrient loss, decreasing soil carbon sequestration and raising CO2 emissions. Using waste instead of fossil fuels may reduce CO2 emissions [68]. Finally, GNE, inflation, fertilizer use, and industrial growth greatly influence CO2 emissions because industrial growth in China often entails expanding manufacturing and output at a high level, increasing CO2 emissions. The excess use of fertilizers to increase agricultural production causes environmental pollution, especially increasing CO2 emissions [21]. The development of targeted industry-based greenhouse gas reduction strategies. The top-down analysis allows the assessment of tourism as a sector within the wider economy [69]. For every 1% increase in tourism demand, foreign direct investment (FDI) has a 0.22% positive effect and a 0.54% negative effect. In China, an unbalanced correlation exists between foreign direct investment (FDI) and tourism, seemingly stemming from a unidirectional causal relationship [70]. While inflation helps to lower CO2 emissions, it also lowers industrial growth when people’s purchasing power declines due to rising inflation. The confirmation of cointegration among the variables and both short- and long-run regression parameters indicated evidence of a U-shaped association between the level of industrial growth and CO2 emissions [62]. 4.1 Descriptive statistics Table 2 demonstrates the descriptive statistics for CO2 emissions, fertilizer consumption (FP) (percentage of output), and fertilizer consumption (FR) (kg per hectare of arable land). Descriptive statistics show that inflation (IN) and gross national expenditures (GNE) achieved their highest levels. The distribution of industrial growth (IG) (annual growth) as a percentage of overall GDP growth is adversely skewed. The study has a sample of data on CO2 emissions, fertilizer consumption (FP), and fertilizer consumption (FR). Download: PPT PowerPoint slide PNG larger image TIFF original image Table 2. Descriptive statistics. https://doi.org/10.1371/journal.pone.0297413.t002 While overall GDP is growing, the industrial sector is growing at a different pace or may even be contracting. The interpretation of the distribution’s means, median and skewness depends on the specific context and the goals of the analysis. In some cases, an adversely skewed distribution may be desirable, such as when trying to reduce economic inequality and more sustainable economic growth. A more symmetrical distribution is preferable, aiming for balanced economic growth across all sectors. Table 3 demonstrates the 1st difference level of IG, FP, and GNE non-stationary. In time series analysis, the factors’ stationarity is essential because the non-stationarity of a unit root in the factors prompts erroneous relapse examinations. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 3. Demonstrates the unit root test results. https://doi.org/10.1371/journal.pone.0297413.t003 The results demonstrate that the data contain a unit root; the IG is non-stationary and may show a trend or drift over time. FP and IN exhibit non-stationary unit roots, indicating temporal drift. GNE data has a trending non-stationary unit root, and the non-stationary ADF test shows a unit root for CO2 emissions. Rising oil prices cut CO2 emissions, limit high-carbon energy use, increase renewable power demand, and lower real GDP [45]. The investigation may proceed to the ARDL and NARDL testing procedures given the ARDL test conditions, no second request coordinated, or I (2) variable is found. An acceptable VAR lag period is necessary before employing the ARDL bound testing approach. Table 4 demonstrates an acceptable lag length for the VAR Lag Order Selection Criteria. Vector Auto-regression (VAR) models have been used to model the dynamic relationship between the study time series variables. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 4. Demonstrates the VAR lag order selection criteria. https://doi.org/10.1371/journal.pone.0297413.t004 The AIC is a measure of the relative quality of a statistical model for the set of data. The BIC is similar to the AIC but places a higher penalty on models with other parameters. The HQIC is a modification of the AIC that provides a more accurate estimate of the optimal lag order for VAR models. The FPE is another measure of the quality of a statistical model that penalizes models with other parameters. 4.2 Wald test parameters Table 5 demonstrates the Wald test results, a statistical test used to determine whether a group of parameters in a regression model are jointly significant. In the context of a VAR model, the Wald test has been used to test the joint significance of the coefficients of multiple lagged values of different time series variables, including Gross National Expenditures (GNE), inflation (IN), fertilizer consumption (FP), and industrial growth (IG), on CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 5. Demonstration of wald test parametric statistical analysis. https://doi.org/10.1371/journal.pone.0297413.t005 The results demonstrated the influence of GNE, FR, IG, IN, and IG on CO2 emissions. The Wald test has been used for the joint significance of coefficients in the ARDL and NARDL models. To perform the Wald test, the study first estimates the ARDL model. The study used the Wald test statistic to calculate the p-value for the test, and the p-value is less than a chosen significance level of 5%. Where C is a vector of the estimated coefficients of the interaction terms, and R is a matrix of the estimated covariance of the coefficients of the interaction terms. According to [46], government expenditures on R&D for renewal energy cause a decrease in CO2 emissions, but due to high expenditures, Chinese companies are still using fossil fuels, which increases the CO2 emissions. The Wald test further demonstrated that ARDL and NARDL models depict relationships between data properties. 4.3 ARDL model and bounds test Table 6 demonstrates the [47] ARDL model in various orders, I (0) and I (1), in the short run. The ARDL model can handle data that are stationary in any order (I(0) or I(1)). The study has applied the [47] Autoregressive Distributed Lag (ARDL) model of order (2, 0, 3, 1, 1, 1) to estimate the short-run and long-run effects of independent variables (FP, FR, IG, IN and GNE) on a dependent variable (CO2 emissions). Download: PPT PowerPoint slide PNG larger image TIFF original image Table 6. Displays the ARDL model estimations for all variables in the short run. https://doi.org/10.1371/journal.pone.0297413.t006 The study has analyzed the coefficients of FP, FR, IG, IN, and GNE influence CO2 emissions. The results look at the sign and magnitude of each coefficient of independent variables FP, IG, IN and GNE statistical significance (using t-tests or p-values) to see whether these variables positively affect CO2 emissions. The mechanism’s effectiveness in promoting sustainable corporate development; however, implementing the mainland-HK Stock Connect has primarily boosted the leading firms [48]. Table 7 demonstrates the ARDL model’s long-run assessments with the bound test. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 7. Demonstration of ARDL model and Bounds Test results in the long run. https://doi.org/10.1371/journal.pone.0297413.t007 The long-run coefficients provide information on the long-term impact of changes in the FP, FR, IG, IN, and GNE influences on CO2 emissions. The long-run assessments with the bound test involve examining the long-run coefficients and statistical significance. A statistically significant coefficient reveals the long-term influence of FP, FR, IG, IN, and GNE on CO2 emissions. The government expenditures on energy technologies that meet the electricity demand, considering economic and environmental parameters [49]. The environment’s quality is influenced over the long run by IG, GNE, FR, and FP impacts on CO2 emissions are significantly and negatively impacted by IN, GNE, and FP, although IN and FP directly correlate to CO2 emissions. The industrial and resource curse concept is supported by a country’s higher GNE and IG and increased loss of natural resources, and China is eventually facing environmental pollution. Most nitrogen is lost to the environment, particularly soil, water, and air, leading to non-point source pollution [50]. Table 8 exhibits ARDL model error correction regression. Error correction evaluates the deviation from the long-run equilibrium relationship between GNE, IG, FR, IN, FP, and CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 8. ARDL model error correction regression. https://doi.org/10.1371/journal.pone.0297413.t008 The results demonstrate that the GNE, IG, FR, IN, and FP are in balance in the long run, but these are out of equilibrium in the short run. The short-run GNE is out of equilibrium, and the long-run and short-run FR & FP discrepancies are resolved within a year as IG lags in adjusting to the independent variables. According to [51], renewable energy positively impacts inflation and lowers CO2 emissions. The CO2 emissions are considerably detrained by the lag between the 1st and 3rd difference in GNE, IG, FR, IN, FP and CO2 emissions, which are significantly correlated. Even if it changed sign in the second and third lags, the short-run demand situation on IG and IN is strong and good. [52] found that moving government funding to public asset investments increases environmental pollution. The short-run impact of the stringency variable is unfavorable and negligible at the 5% significance level. 4.4 ARDL and NARDL models and F-bound test Table 9 asserts the presence of a long-run cointegration relationship by validating the bound test estimations for both ARDL and NARDL. The F-Bounds test has been applied to confirm a cointegration relationship of independent and dependent variables in the long run. The F-Bounds test is a statistical test used to determine whether cointegration exists between all the variables in an ARDL model. The F-statistic is then compared to the distribution’s upper and lower critical values, which are based on the number of variables in the model and the sample size. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 9. The f-Bounds Test confirmed a cointegration connection between the variables. https://doi.org/10.1371/journal.pone.0297413.t009 Table 10 demonstrates the NARDL model results for all the variables in the short run. The study has applied the NARDL model of order (2, 0, 0, 3, 0, 0, 0, 2, 2, 1, 0) to determine the results of all variables. It is helpful to analyze the short-run consequences of changes in emissions of GNE, IG, FR, IN, and FP using the NARDL model [53]. In the short run, the NARDL model can estimate a shock’s immediate and lagged effects on GNE, IG, FR, IN, FP and CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 10. Dynamic estimation of the NARDL Model in the short-run. https://doi.org/10.1371/journal.pone.0297413.t010 The results of the NARDL model in the short run will depend on the specific variables and data used in the analysis. The results show that the GNE, IG, FR, IN, and FP effects positively impact CO2 emissions. Inflation (IN), which is large and positive, has a considerable short-run impact on rising CO2 emissions. Between 0.03% and 0.97% of China’s GDP is used to offset the cost of inflation, with foreigners and investors bearing most of the burden [54]. Although China has agreed on the need to reduce CO2 emissions completely, there are still disparities in regional emissions. Reduced carbon emissions serve the public benefit and reveal a significant positive externality that is challenging to address in the market [11]. Due to China’s and India’s industrial revolutions, industrial growth (IG) is the primary factor driving the country’s rise in CO2 emissions in the next decades [55]. FP and FR impact short-term CO2 emissions and positively correlate with other parameters. Spreading information to fertilizer wholesalers, crop advisors, farmers, and agricultural and environmental authorities should boost BMP consumption [56]. Positive shocks in FP have a positive and substantial coefficient that greatly influences CO2 emissions. An increase in FP causes CO2 emissions to rise, whereas negative shocks to IN usage substantially impact CO2 emissions. A decrease in the usage of renewable energy sources leads to a rise in CO2 emissions, according to the negative IN coefficient. Table 11 reveals that the NARDL model long-run coefficient will converge to equilibrium. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 11. NARDL model long-run and limited constant with no trends. https://doi.org/10.1371/journal.pone.0297413.t011 The results demonstrate that the positive shocks to GNE and IG meaningfully affect fossil fuel byproducts’ high likelihood worth and negative coefficient esteem. China’s CO2 emanations are increasing over the long haul because the nation is making an ever-increasing number of farming items [57]. A positive change in IG affects CO2 discharges, showing that contamination in the climate decreases as IG improves. It is another significant component affecting fossil fuel byproducts when a nation’s funds deteriorate; because less cash is in the in the evolving hands, and pollution levels go up. Conversely, negative shocks to IG show a reasonable connection with fossil fuel byproducts. 4.5 NARDL bounds test Table 12 demonstrates the F-Bounds test to assert the cointegration connection in the short- and long-run. The NARDL model with the Bounds test has been used in econometric analysis to determine the relationship between study macroeconomic variables and CO2 emissions. Once the NARDL model is estimated, the study will conduct the F-Bounds test to assess the existence of cointegration between the variables. The F-Bounds test involves calculating the F-statistics upper and lower bounds based on macroeconomic variables’ impact on CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 12. F-Bounds test to assert the cointegration connections. https://doi.org/10.1371/journal.pone.0297413.t012 Industrial revolutions have diverse implications for achieving net-zero carbon emissions [58]. The study assessed the cointegration connection between the variables and gained insights into their long-run relationships by calculating the upper and lower bounds of the F-statistic. China’s industrial sector investment volume increases CO2 emissions [34]. 4.6 CUSUM and CUSUM of square graphs Fig 2 demonstrates the CUSUM and CUSUMSQ of the square model’s stability. The CUSUM test involves calculating a sequence of test statistics that represent the cumulative sum of the residuals from the regression model, and the results of test statistics have been plotted against the sample size and a critical value. The CUSUMSQ test involves calculating a sequence of test statistics representing the cumulative sum of the squared residuals from the regression model. The parameter stability is evaluated using the cumulative sum of recursive residuals (CUSUM) and cumulative sum of squares (CUSUMSQ) tests [59]. The study used the CUSUM and CUSUM of Squares (CUSUMSQ) tests to discover the NARDL model residuals and the link between GNE, IN, IG, FP, and CO2 emissions. Study test data were plotted against sample size and a critical value. CUSUM and CUSUM of Square tests identify time series mean shifts. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 2. The CUSUM and CUSUM of squares test results of macroeconomic variables. Source: Author’s illustration. https://doi.org/10.1371/journal.pone.0297413.g002 The graphs indicate a significant change in the mean of FP, GNE, IN, and IG related to CO2 emissions, and it suggests a potential causal relationship between FP, GNE, IN, IG and CO2 emissions. Table 13 demonstrates the diagnostic inspection test for verification of the ARDL and NARDL model results. The ARDL and NARDL models have assumed that the errors are normally distributed, and deviations from normality can lead to biased coefficient estimates [60]. The heteroskedasticity-hearty Breusch-Agnostic test is more reliable than the "wild bootstrap form" of the standardized unique Breusch-Agnostic test [61]. The diagnostic tests have been applied to identify potential problems with the ARDL and NARDL models and improve the overall performance of all the variables. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 13. Diagnostic test inspection. https://doi.org/10.1371/journal.pone.0297413.t013 The Jarque-Bera test’s probability values are more than 5%, which shows that the null hypothesis of normality is not disproved. The evidence for a U-shaped connection between the rate of industrial expansion and CO2 emission was supported by both short-run and long-run regression parameters [62]. The hidden ARDL and NARDL models’ relapses match very well; both are critical at the 1% level worldwide. These models passed the Lagrange Multipliers (LM) test for heteroskedasticity, the Jarque-Bera test, the Breusch-Agnostic Godfrey test, and the RESET test. The consumption of fertilizers and fossil fuels consumption increased environmental pollution [63]. 4.7 Estimation of a ramsey RESET model Table 14 demonstrates the RESET test confirming the results of the ARDL and NARDL models. The Ramsey RESET test is a diagnostic test applied to determine a nonlinear regression between the independent and dependent variables. The Ramsey RESET test is also a diagnostic tool used with other diagnostic tests to ensure the model is correctly specified. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 14. Ramsey RESET test with F-test summary. https://doi.org/10.1371/journal.pone.0297413.t014 It estimates the regression model of CO2 emissions as a function of Gross National Expenditures (GNE), Inflation and Industrial Growth. The results of the Ramsey RESET test have been interpreted in conjunction with other diagnostic tests and considerations, and the residuals for patterns and testing for multi-collinearity among the regressors. 4.8 Granger causality test Table 15 demonstrates the Granger causality scales between the components after the cointegration coefficients. The Granger causality test is a statistical technique used to determine whether one variable is used to predict another variable. The study has applied the Granger causality test to analyze the time series data for stationarity. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 15. Demonstrates that the Granger causality test estimations. https://doi.org/10.1371/journal.pone.0297413.t015 The results reveal that the positive shocks of GNE to CO2 emissions are correlated in both directions; negative shocks in IG, FP, and GNE Granger cause CO2 emissions. A unidirectional association between IG usage and FR confirms China’s CO2 emissions. CO2 emissions have positive shocks in IG, negative shocks in IN, and negative shocks in FP, which are all brought on by Granger. To validate the GNE theory, there is also a unidirectional association between FR, IN, GNE, and FP. Table 16 demonstrates the Granger causality test result for the effectiveness of all variables on CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 16. Granger causality test result for the effectiveness of CO2 emissions. https://doi.org/10.1371/journal.pone.0297413.t016 The results of the Granger causality test for the association between GNE, IG, and FR are intriguing. These results show that neither the Granger causality estimation from GNE to IN nor IN to IG is statistically significant. The NARDL model reveals that changes in energy usage, fertilizer use, and agricultural carbon emission lead to changes in cereal food production, both positively and negatively (Koondhar et al., 2021) [64]. The findings demonstrate that GNE, FP, IG, and FR growth will only increase CO2 emissions. Table 17 shows the impulse response function of CO2 emissions to confirm the relationship validity between the study variables. The impulse response function indicates that economic activity raises CO2 emissions mathematically. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 17. Demonstrates the Impulse Response Function of CO2 emissions. https://doi.org/10.1371/journal.pone.0297413.t017 The results demonstrate that CO2 emissions have an inverse connection with negative shocks to GNE and IG. The positive shocks to IN decrease CO2 emissions, whereas an increase in IG and GNE also increases the CO2 emissions. China’s CO2 emissions rise in response to positive shocks to FP and FR. Within a decade, green finance will significantly reduce fertilizer use and agricultural carbon emissions [65]. CO2 emissions will fall along with an increase in FP, and China’s GNE and IG will also significantly influence CO2 emissions. 4.9 Variance decomposition analysis Variance decomposition analysis has been applied to decompose the variance of the macroeconomic research variables and CO2 emissions. A quarter of all human CO2 emissions are brought on by land usage and agricultural output [66]. There are positive shocks of IG, FR, and GNE to CO2 emissions, but FP has a negative shock FP to CO2 emissions. Table 18 demonstrates the variance decomposition analysis (VDA) of CO2 emissions. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 18. Variance decomposition analysis of CO2 emissions. https://doi.org/10.1371/journal.pone.0297413.t018 The results demonstrate that FP, GNE, IG and IN significantly impact CO2 emissions. The GNE is the most important variable, and policymakers might focus on promoting economic growth through sustainable and low-carbon technologies while reducing energy consumption and improving energy efficiency. 4.10 Discussions The findings show that greater gross national expenditures (GNE) correspond to greater economic activity, which increases CO2 emissions. Currently, no regulatory interactions link increasing economic development and reducing carbon dioxide emissions at the national level [67]. Most economic activity depends on energy usage, normally produced using fossil fuels that emit CO2. Higher inflation rates cause lower CO2 emissions. Using and manufacturing fertilizers have expanded, increasing greenhouse gas emissions. In addition, excessive fertilizer usage can result in soil deterioration and nutrient loss, decreasing soil carbon sequestration and raising CO2 emissions. Using waste instead of fossil fuels may reduce CO2 emissions [68]. Finally, GNE, inflation, fertilizer use, and industrial growth greatly influence CO2 emissions because industrial growth in China often entails expanding manufacturing and output at a high level, increasing CO2 emissions. The excess use of fertilizers to increase agricultural production causes environmental pollution, especially increasing CO2 emissions [21]. The development of targeted industry-based greenhouse gas reduction strategies. The top-down analysis allows the assessment of tourism as a sector within the wider economy [69]. For every 1% increase in tourism demand, foreign direct investment (FDI) has a 0.22% positive effect and a 0.54% negative effect. In China, an unbalanced correlation exists between foreign direct investment (FDI) and tourism, seemingly stemming from a unidirectional causal relationship [70]. While inflation helps to lower CO2 emissions, it also lowers industrial growth when people’s purchasing power declines due to rising inflation. The confirmation of cointegration among the variables and both short- and long-run regression parameters indicated evidence of a U-shaped association between the level of industrial growth and CO2 emissions [62]. 5. Conclusions and recommendations The rapid economic growth and development have led to an increased reliance on fossil fuels, particularly oil, which has significantly increased CO2 emissions in China. The study has been researched to determine the influences of gross national expenditures (GNE), inflation (IN), fertilizer consumption (FP), and industrial growth (IG) on CO2 emissions using time series data from 1960 to 2022. The methodology has applied ARDL and NARDL models to analyse the short- and long-run data. The Granger causality, IRF, and VDA are also utilized to determine the relationship between the GNE, IN, FP, IG, and CO2 emissions. The F-bound test has been used to confirm the long-run cointegration of all the variables. The results demonstrate that momentary CO2 emissions have a solid and unfavorable relationship with GNE and IG. The Granger causality results show that FR, IG, FP, and GNE significantly impact CO2 emissions. The relationship between inflation and GDP is positive, but FP, GNE, and IG hurt CO2 emissions. Additionally, it is important to consider other factors that may influence CO2 emissions, such as population growth, economic conditions, and energy policies. The IRF found negative shocks to GNE and CO2 emissions but positive shocks to GNE, IG, and CO2 emissions. A negative shock to GNE is thought to result in a reduction in CO2 emissions, according to the IRF research, which also revealed negative shocks to GNE and CO2 emissions. A positive shock to GNE or IG is thought to boost CO2 emissions, according to the IRF analysis, which also discovered positive shocks to GNE, IG, and CO2 emissions. The VDA shows negative IG, GNE, FR, and CO2 emissions shocks. The VDA analysis reveals negative IG, GNE, FR, and CO2 emissions shocks. The possible links between the four variables investigated in the VDA and other variables could affect the outcome variables. The impact of FR on CO2 emissions has altered due to institutional sufficiency. The FP and IN positively impact CO2 emissions, and a 1% increase in GNE will also increase CO2 emissions. Higher industrial growth (IG) has an unequal impact on CO2 emissions. The impact of industrial growth on CO2 emissions may depend on the energy intensity of the industrial sector, the use of renewable or fossil fuel-based energy sources, the efficiency of production processes, the level of technology and innovation, and the environmental regulations and policies in place. Because of the rise in fossil fuel byproducts caused by FR and FP, expansion (IN) also exhibits unbalanced behavior. FP, IG, and GNE are key drivers of the increase in CO2 emissions in China, as the burning of fossil fuels is the largest contributor to anthropogenic CO2 emissions. As the demand for energy continues to grow, particularly in developing countries, the use of fossil fuels is likely to increase, leading to further increases in CO2 emissions. China is the world’s largest emitter of CO2 emissions, accounting for over a quarter of global emissions. FP, IG, and GNE are identified as key drivers of the increase in CO2 emissions in China. CO2 emissions increase global temperatures, precipitation patterns, extreme weather events, and ocean levels. There is a need to shift towards cleaner and more sustainable sources of energy, such as renewable energy sources like solar, wind, and hydropower. 5.1 Future research suggestions Future research can investigate how industrial growth affects emissions across different industries or countries and identify policies or practices that can help reduce emissions associated with industrial growth. Overall, these research suggestions could help better understand the complex relationships between economic growth, agricultural practices, industrial development, and CO2 emissions and identify strategies to reduce emissions and mitigate the impacts of climate change. 5.1 Future research suggestions Future research can investigate how industrial growth affects emissions across different industries or countries and identify policies or practices that can help reduce emissions associated with industrial growth. Overall, these research suggestions could help better understand the complex relationships between economic growth, agricultural practices, industrial development, and CO2 emissions and identify strategies to reduce emissions and mitigate the impacts of climate change.
TI - The hidden costs of inflation: A critical analysis of industrial development and environmental consequences
JO - PLoS ONE
DO - 10.1371/journal.pone.0297413
DA - 2024-08-05
UR - https://www.deepdyve.com/lp/public-library-of-science-plos-journal/the-hidden-costs-of-inflation-a-critical-analysis-of-industrial-gGsdYxiyZt
SP - e0297413
VL - 19
IS - 8
DP - DeepDyve
ER -