The climate change crisis requires strong, rapid and sustained global actions to limit greenhouse gas emissions in the coming decades. Although most developed countries are in the emissions reduction stage on the path towards carbon neutrality, numerous developing countries, mainly distributed along the Belt and Road, are facing severe pressure in the low-carbon transition. Therefore, a large amount of financial support from developed countries is required. Considering that the Belt and Road countries receive over 60% of global climate aid, with energy as the largest usage sector, the present study evaluated the effects of climate aid on carbon emissions and the impact mechanism through energy structure optimization. A systematic panel approach using system generalized method of moments estimators was established, and then the panel data of 93 Belt and Road countries from 2000–2018 were used for empirical analysis. The results show that climate aid has a significant reduction effect on the carbon emissions intensity of Belt and Road countries, and the dominant component of climate aid, i.e., mitigation aid, shows better carbon emissions reduction benefits than adaptation aid. The impact mechanism clarifies that climate aid has dual carbon emissions reduction effects in BRI countries with an intermediate energy structure, indicating that climate aid not only directly reduces carbon emissions by increasing carbon reduction resources in BRI countries but also indirectly reduces carbon emissions by promoting renewable energy to optimize the energy structure. This paper suggests that to promote the low-carbon transition and carbon neutrality process in Belt and Road countries, more climate finance support should be provided by developed countries as pledged, the effectiveness of climate aid should be improved by investing in projects with significant carbon reduction effects such as renewable energy, and monitoring and performance evaluation of climate aid should be strengthened.
Research Article
Does Climate Aid Truly Work for Carbon Emissions Reduction? Evidence from the Belt and Road Countries
https://doi.org/10.21203/rs.3.rs-1419811/v1
This work is licensed under a CC BY 4.0 License
You are reading this latest preprint version
Climate aid
Climate finance
Carbon emissions
Mediating effect
Renewable energy
Carbon neutrality
Climate change
Belt and road initiative
Climate change is one of the most severe challenges humankind has ever faced. To manage the impacts and minimize the risks, strong, rapid and sustained global actions are required to limit cumulative CO2 emissions in the coming decades, reaching at least net zero CO2 emissions, along with strong reductions in other greenhouse gas (GHG) emissions (IPCC 2021). An increasing number of countries are adopting carbon neutrality as a national strategy and making long-term low-carbon emissions plans. By the end of 2021, 136 countries and regions had proposed or were preparing to propose carbon neutrality goals, covering 88% of GHG emissions, 90% of Gross Domestic Product (GDP) and 85% of the population worldwide. However, economic development and the carbon emissions status and trends vary greatly among countries. Most developed countries have achieved carbon dioxide peaking and are in the emissions reduction stage on the path towards carbon neutrality, while numerous developing countries are in the carbon emissions growth stage or plateau stage or have not yet started industrialization, bringing stringent pressure and uncertainty to the achievement of global carbon neutrality. As most of the latter category is distributed along the Belt and Road, more attention should be given to how to promote the low-carbon transition of these countries.
Since the launch of the Belt and Road Initiative (BRI) in 2013, 146 countries have become signatories, including 46 in Africa, 44 in Asia, 26 in Europe, 19 in the Americas and 11 in Oceania. From 2013 to 2019, the GDP of BRI countries (including China) increased from $26,024 billion to $33,655 billion, with the percentage of global GDP increasing from 36.86–39.78%. Moreover, BRI countries are the main source of global carbon emissions, driving the increase in global carbon emissions from 56.86% in 2013 to 58.72% in 2019 (Fig. 1a) (OWD 2021, WDI 2021). Although the growth rate of carbon emissions in BRI countries is lower than that of GDP, the share of carbon emissions of BRI countries is much higher than that of GDP, and the share of fossil fuels in primary energy consumption is over 90%, indicating economic development in the mode of "high growth, high emissions, and high energy consumption" (Fig. 1b) (EIA 2021). Therefore, on the one hand, it is very difficult for BRI countries to achieve low-carbon transitions and the carbon emissions reduction targets they propose. On the other hand, BRI countries usually have fragile ecological environments and insufficient environmental capacity and are vulnerable to climate change risks, making their low-carbon transitions much more urgent.
Generally, the low-carbon transition paths of BRI countries include industrial transformation, energy transformation, energy saving and efficiency improvement, technological innovation, and low carbon consumption. As energy consumption is the main driver of carbon emissions, low carbonization of the energy structure is vital to achieve carbon neutrality, in which the development of a high proportion of renewable energy is key. It is estimated that the global share of renewable energy generation will increase from 29% in 2020 to 70% in 2050, with solar PV and wind power accounting for approximately 50% (IRENA 2018). BRI countries have advantages for the development of renewable energy. Many BRI countries have excellent light radiation and wind resource conditions. For example, the light intensity in Southeast Asia and the Middle East and the wind density in Central Asia, Eastern Europe, and South America are both higher than the global average of existing projects. In addition, the running hours of renewable energy projects in the operation of BRI countries are relatively high compared to the global average. Moreover, the growth rate of energy demand will be much higher than the global average, there will be sufficient land resources and resource distribution, and load centres can be matched geographically. It is estimated that the total renewable energy investment potential is $678 billion in BRI countries (Munoz et al. 2018). However, BRI countries have limited finances for renewable energy and other climate-related areas, and without the strong support of developed countries, it is difficult to implement a low-carbon transition and achieve net zero emissions.
To support developing countries in addressing climate change, the United Nations Framework Convention on Climate Change (UNFCCC) has established a climate finance mechanism since 1997 (Charlene Watson 2021, UN 1997). As the main recipients of international climate finance, BRI countries have received climate finance support through various bilateral and multilateral channels, with a total amount of 71,485 climate aid projects and a total funding of over $322 billion from 2000 to 2019, which accounts for over 60% of the global total (over $535.4 billion) (OECD 2021). In regard to time, climate aid funding was only a few billion USD before 2009, and after the Copenhagen Accord of 2009, in which developed countries pledged to raise $100 billion per year by 2020, the scale of funding increased significantly, exceeding $50 billion in 2019. In regard to finance type, climate aid is dominated by mitigation finance of approximately $229.1 billion (71.1%) and adaptation finance of approximately $121.6 billion (37.7%) (Fig. 2). In regard to finance usage, $91.7 billion is distributed to energy, accounting for approximately 28.5% of total climate aid, versus $50.6, $38, and $36.5 billion for transport and storage, water supply and sanitation, agriculture, forestry and fishing, respectively. As BRI countries receive a large amount of climate aid and a large portion is directed towards energy, it is sensible to examine whether climate aid has effects on carbon emissions reduction and the role that energy plays.
Therefore, this paper aims to examine the impact of climate aid on carbon emissions in BRI countries, to clarify the mechanism of climate aid and energy structure on carbon emissions in BRI countries, and to provide a theoretical and empirical foundation for better utilizing climate aid and building low-carbon energy systems. For these, we theoretically proposed a dynamic panel data model to explore the relationship between climate aid and carbon emissions. We further analyzed the carbon emissions reduction mechanism and empirically explored the relationship between climate aid, energy structure and carbon emissions by using a mediation model. The rest of the paper is organised as follows. In Section 2, we review the literature on the impact of climate finance on carbon emissions and the emissions reduction mechanism. In Section 3, we introduce the empirical model and describe the data. In Sections 4 and 5, we discuss the results, and in Section 6, we summarise with policy implications.
2.1 Carbon emissions-driving factors and emissions reduction paths in BRI countries
There are extensive studies on carbon emissions-driving factors in BRI countries that indicate the key role of energy. (Fan et al. 2019) analysed the contributions of seven driving factors of the total carbon dioxide emissions changes for BRI countries from 2000 to 2014 and indicated that economic development and energy consumption are the two most important factors affecting carbon emissions growth, while carbon abatement technology changes and potential energy intensity changes are the main inhibiting factors. (Zhu and Gao 2019) measured the CO2 emissions and driving factors of the transportation industry in BRI countries from 2005 to 2015 by using a panel data model and showed that the per capita GDP, urbanization level, and energy consumption structure have positive effects on carbon emissions, while the technology level and trade openness have negative effects. (Han et al. 2020) compared carbon emissions among countries in and outside the Belt and Road area from 1990 to 2015 and found that most of the Belt and Road regions achieved a rapid GDP growth rate with increasing carbon emissions, and the investment-driven type of emissions demonstrated significant growth. Therefore, greening infrastructure investments in BRI countries will have a large impact on carbon emissions reduction globally (Yang and Yang 2019). Among the emissions reduction paths, reducing fossil fuel dominance and expanding renewable energy in electric power development is of great importance and urgency (Tao et al. 2020). (Gu and Zhou 2020) analysed China's investments in wind and solar energy projects and the carbon emissions reduction benefits to the host country, and the results showed that the 36 renewable energy investment projects with an installed capacity of 15.75 GW in BRI countries achieve at least 48.69 MtCO2 emissions reduction. (Zhang et al. 2019) pointed out that power generation efficiency is the essential factor to improve carbon emissions performance in BRI countries and that promoting clean and efficient thermal power is a pragmatic priority for green power development and emissions reduction paths for BRI countries, optimizing energy-related infrastructure investment as one of the most important reduction paths.
2.2 Climate aid and carbon emissions
The relation between climate aid and carbon emissions has gradually drawn the attention of academics, but conclusions usually vary because of variable selection, model setting, and other aspects (Ellis et al. 2013). Some studies have pointed out that climate aid has a positive effect on carbon emissions reduction. (Wu et al. 2015) established the climate financing model RICE and analysed the impact of several financing program scenarios. The results showed that sustained climate financing can help to combat global climate change. (Carfora et al. 2017) analysed the relationship between climate finance and GHG emissions and found that funds directed to “fast-start finance” met the requirements of ensuring the reduction in GHG and promoting the sustainable development of developing countries. (Carfora and Scandurra 2019) evaluated the effects of climate funds received by recipients and found that the policy contributed to the decrease in GHG emissions and promoted the replacement of fossil sources with renewable sources. However, some studies did not find a significant impact of climate aid on carbon emissions reduction. (Kretschmer et al. 2013) examined the impact of total aid, industrial sector aid and energy sector aid on carbon emissions in developing countries and concluded that total aid and industrial sector aid had little impact on carbon emissions in recipient countries, while energy aid helped reduce energy intensity but not carbon emissions. (Bhattacharyya et al. 2018) performed an empirical audit of the effectiveness of energy-related aid in tackling CO2 emissions and showed that aid had no significant CO2 emissions reduction effects. (Chung et al. 2018) examined the effects of climate technology-related development finance on CO₂ emissions of the recipient country and showed that climate technology-related development finance did not have a significant effect on CO₂ emissions as a whole, but different types of technical aid had heterogeneous effects in different sectors.
2.3 Impact mechanism of climate aid on carbon emissions
Existing studies explored the impact mechanism of climate aid on carbon emissions from two aspects, namely, environmental economic theory and impact path. In the early stage, many studies focused on environmental economic theory. Economists try to examine the mechanism from the perspectives of environmental policy (Copeland and Taylor 1994), abatement efficiency (Ono 1998), or the supply and demand of polluting goods (Naito 2003). Recent studies have gradually addressed the impact path from the perspective of energy structure, economic scale and structure, industrial structure, technology innovation, etc. For example, (Wang and She 2018) divided the impact of climate aid on carbon dioxide emissions in China into direct effects and indirect effects (consisting of scale effects, structure effects and technology effects) and then established a dynamic optimization model for empirical analysis. (She and Wu 2019) found that the technology advancement of recipient countries cannot only reduce domestic carbon emissions but also help strengthen the carbon reduction effect of energy-related aid. (Shen et al. 2021) examined the long- and short-term impacts of several variables on carbon emissions and found that green investment is negatively linked to CO2, indicating the promotion of green investment to control carbon emissions.
Specifically, an increasing number of studies have examined the impact mechanism through energy-related optimization because of the close relationship between energy consumption and carbon emissions. (Carfora and Scandurra 2019) showed that the effectiveness mechanism of climate aid on carbon emissions reduction lies in promoting renewable energy source generation. (Wu et al. 2021) investigated the influencing mechanism of climate aid based on the relations among climate aid, energy structure and carbon emissions in recipient countries with the mediation model. The results showed that climate aid has not only direct carbon reduction effects but also indirect carbon reduction effects through mediation of the renewable energy structure. (Liobikienė and Butkus 2019) also used renewable energy as a mediation variable to investigate the channels through which economic factors affect GHG emissions. The results showed that GDP, urbanization, and trade contributed to reducing carbon emissions only via energy efficiency but not via renewable energy.
In conclusion, existing studies have investigated the nexus of climate aid and carbon emissions for different countries and regions, but few studies have taken BRI countries as the study object. In addition, the impact mechanism with an intermediate energy structure needs to be verified for BRI countries. Moreover, climate aid includes two types, namely, mitigation and adaptation, but existing studies did not distinguish between the mitigation and adaptation components of relevant climate aid, which provides another aspect for further research. Finally, most panel data are heterogeneous and non-stationary, and the issues of structural breaks, cross-sectional dependence or autocorrelation between variables must be considered to establish an empirical model. Therefore, this study attempts to conduct a systematic empirical analysis of the relationship between climate aid, energy structure and carbon emissions in BRI countries to explore the effects of climate aid on carbon emissions in these countries and the impact mechanism through the energy structure. This study focuses on both climate aid and its components of mitigation and adaptation climate aid, establishes a systematic panel approach using system generalized method of moments (GMM) estimators, and empirically investigates the effects of climate aid on carbon emissions and the mediation of energy structure between climate aid and carbon emissions in BRI countries.
3.1 Data
This study explores the impact of climate aid on carbon emissions in BRI countries from 2000 to 2018. Considering data availability, 93 BRI countries that received climate aid (see Appendix A) are selected for the empirical analysis. The dependent variable, CO2 emissions intensity, is calculated as per unit of GDP CO2 emissions, which was recently used by (Wu et al. 2021). The key independent variable, climate finance, is the climate aid that BRI countries receive on a commitment basis. We use three proxies for climate finance, namely, total climate finance and its two components, mitigation climate finance and adaptation climate finance. Based on the existing literature, we control for a set of variables to capture the influencing factors of carbon emissions, including economic growth, population, research and development (R&D), trade openness, urbanization and industrialization. Specifically, economic growth is measured by GDP per capita (constant 2010 US$), and urbanization is denoted as the share of the urban population in the total population (Bosah et al. 2021, Hussain et al. 2021). According to (Omri and Bel Hadj 2020), trade openness is expressed by total exports plus imports as a share of GDP. Following (Hu et al. 2020), we use the proportion of industrial added value to GDP to measure industrialization. R&D is measured by scientific and technical journal articles. CO2 emissions data are sourced from Our World in Data, a project of the non-profit organization Global Change Data Lab based in the United Kingdom. This database is continuously updated with the best available data so that it is trusted in research and media such as Science, Nature, PNAS, the BBC, and CNN(OWD 2021). The climate finance data are sourced from the OECD Development Assistance Committee (DAC) Creditor Reporting System (CRS) database, which includes the most comprehensive, publicly available, activity-level data on climate-related development finance since the year 2000 (OECD 2021). The data of all the control variables are sourced from the World Development Indicators (WDI) database of the World Bank (WDI 2021). Appendix B provides detailed definitions and data sources of all variables.
3.2 Methodology
3.2.1 Baseline model
The panel empirical model investigates the impact of climate aid on CO2 emissions in BRI countries. As mentioned previously, the baseline model is given as follows:
$$\ln C{E_{it}}={\alpha _0}+{\alpha _1}\ln C{F_{it}}+\sum {{\alpha _j}{X_{jit}}} +{\mu _i}+{\nu _t}+{\varepsilon _{it}}$$ 1
where α0 to αj are coefficients to be estimated. lnCEit is CO2 emissions intensity in the ith BRI country at time t. lnCFit represents climate aid received by a BRI country, which can be expressed by three proxies, namely, total climate finance and its two components, mitigation climate finance and adaptation climate finance. The vector Xjit comprises the control variables involved in this study, which are described in the previous Data section. µi and νt represent the fixed effects for individual and time, respectively. εit represents stochastic disturbance.
The system GMM method is chosen for the estimation. GMM was proposed by (Arellano and Bond 1991), and system GMM was proposed by (Blundell and Bond 1995). There are four key reasons for utilizing the system GMM estimator within the framework of the panel data methodology. First, the GMM technique allows us to address endogenous issues related to the research variables, and system GMM can further handle differences and grouping of equations on a horizontal level. The tools specified for the difference equation are the variable delay value in the level. Second, the variables studied are horizontal equations and first-order difference instruments. The equations also use the generalized moment estimation method. The Monte Carlo simulation by (Blundell and Bond 1998) shows that the system GMM model is the most effective in estimating this dilemma. Third, compared to estimations by the difference GMM estimator, estimations by the system GMM estimator are more efficient, allowing for more instruments through an additional assumption, in which first differences of instruments are uncorrelated with the fixed effects. Finally, compared to a one-step GMM estimator, a two-step GMM estimator tends to be biased downwards with a finite sample and is more efficient. Therefore, two-step system GMM was selected for the estimation. To confirm our research expectation, the overdocumentation test was replaced by the Sargan test with the Hansen test, and the sequence correlation test of Arellano and Bond was also used.
The GMM equation with climate finance and carbon emissions as dependent variables can be written as follows:
$$\begin{gathered} \ln C{E_{it}}={\beta _0}+{\beta _1}\ln C{E_{i,t - 1}}{\text{+}}{\beta _2}\ln (1+C{F_{it}})+{\beta _3}\ln PGD{P_{it}}+{\beta _4}{(\ln PGD{P_{it}})^2}+{\beta _5}\ln PO{P_{it}} \hfill \\ +{\beta _6}\ln R{D_{it}}+{\beta _7}TRAD{E_{it}}+{\beta _8}URBA{N_{it}}+{\beta _9}IND{U_{it}}+{\mu _i}+{\nu _t}+{\varepsilon _{it}} \hfill \\ \end{gathered}$$ 2
where β0 to β9 are coefficients to be estimated. lnCEi,t−1 is the lagged CO2 emissions intensity. The key dependent variable and all the control variables are described in the previous Data section. GDP per capita square is utilized to estimate the EKC hypothesis of BRI countries. In addition, all the variables with absolute values are measured in logarithmic form to eliminate the effects of heteroscedasticity that outliers may cause, while the variables with relative values (TRADE, URBAN, INDU) are measured by their original data. As there is zero value for climate finance, the ln(1 + CFit) form is used.
|
Variable |
Description |
Symbol |
|---|---|---|
|
Carbon dioxide emissions |
CO2 emissions intensity (ton/million USD) |
CE |
|
Climate finance |
Climate-related development finance |
CF |
|
|
|
|
Economic growth |
Gross domestic product per capita |
PGDP |
|
Population |
Total population |
POP |
|
R&D |
Scientific and technical journal articles |
RD |
|
Trade openness |
Sum of exports and imports as a percentage of GDP |
TRADE |
|
Urbanization |
Urban population % of total population |
URBAN |
|
Industrialization |
Industry value added as a percentage of GDP |
INDU |
3.2.2 Mediation model
Climate aid can reduce carbon emissions by optimizing the energy structure in BRI countries. As energy consumption is the main driver of carbon emissions, it has been widely recognized that energy structure optimization is conducive to carbon reduction (Khan et al. 2021, Lin and Ahmad 2017, Shen et al. 2021, Wang et al. 2016, Yu et al. 2018). Energy structure optimization can reduce carbon emissions in two ways: by increasing the proportion of renewable energy and substituting fossil fuels and by promoting the low-carbon transformation of traditional fossil fuel energy. Therefore, climate aid exerts a positive impact on carbon emissions reduction by optimizing the energy structure in BRI countries. That is, the positive transmission mechanism of “climate aid → optimize energy structure → reduce carbon emissions” exists steadily.
Based on the above theoretical analysis, this paper describes the impact mechanism of climate aid on carbon emissions in BRI countries. According to related studies (Pei et al. 2019, Qu and Luo 2021, Wu et al. 2021), the recursive Eqs. (3), (4) and (5) are set to analyse whether the transmission mechanism of intermediary variables exists to more clearly describe the impact path of climate aid on carbon emissions reduction. Following the method proposed by (Wen and Ye 2014), the test of intermediate variables is also used to verify the positive impact on carbon emissions reduction by optimizing the energy structure.
Eq. (3) examines the total effect of climate aid on carbon emissions in the absence of a mediating variable, and coefficient c represents the total effect. Eq. (4) analyses the impact of climate aid on the mediating variable. Eq. (5) investigates the effect of climate aid on carbon emissions by adding the mediating variable. Coefficient c’ represents the direct effect, and ab represents the indirect effect; thus, the total effect is the sum of the direct effect and indirect effect, i.e., c = c’+ ab.
4.1 Panel unit root tests
When using panel data, it is critical to test the relationships between the variables under consideration. We conducted three types of unit root tests, i.e., LLC, IPS, and Fisher–type ADF, for the independent variables and all of our dependent variables from two dimensions, i.e., intercept, intercept and trend. The LLC test assumes that there is a common unit root process across cross-sectional units (Levin et al. 2002), while the IPS test assumes that all cross-sectional units converge towards the equilibrium value at different speeds under the alternative hypothesis (Im et al. 2003). Fisher-type ADF is based on heterogeneous cross-sectional formation (Choi 2001). The test results shown in Table 2 indicate that some variables are not stationary in level order, while the scores for LLC (Δ), IPS (Δ) and ADF (Δ) show highly significant outcomes for all the study variables. The hypothesis of the series containing a unit root can be confirmed at the 1% significance level.
| Intercept | Intercept & Trend | |||||
|---|---|---|---|---|---|---|
| LLC | IPS | ADF | LLC | IPS | ADF | |
| Lnce | -4.7344*** | 0.8947 | 187.2089 | -9.4566*** | -5.4353*** | 235.3485*** |
| D.Lnce | -24.2233*** | -7.6291*** | 317.4974*** | -25.1312*** | -21.4627*** | 235.8888*** |
| Lntcf | -8.3805*** | -14.9159*** | 283.2990*** | -18.8703*** | -15.7653*** | 292.8551*** |
| D.Lntcf | -24.1115*** | -41.6645*** | 596.2804*** | -26.7046*** | -30.6847*** | 451.0505*** |
| Lnadcf | -0.7594 | -5.6157*** | 127.4454 | -6.9661*** | -10.7241*** | 176.7897 |
| D. Lnadcf | -35.9229*** | 213.0745* | 498.7305*** | -29.6350*** | --31.88633*** | 371.3386*** |
| Lnmtcf | -7.2381*** | -16.0747*** | 314.7700*** | -18.2531*** | -15.4535*** | 293.5408*** |
| D. Lnmtcf | -26.8469*** | -42.8955*** | 583.7827*** | -26.4635*** | -35.8393*** | 441.7040*** |
| Lnpgdp | -5.5431*** | 1.5408 | 258.6006*** | -5.5160*** | 1.7060 | 169.2874 |
| D.Lnpgdp | -14.7438*** | -16.0466*** | 243.8185*** | -19.6506*** | -14.4285*** | 269.3768*** |
| Trade | -6.0695*** | -3.8985*** | 255.9128*** | -8.0888*** | -2.8284*** | 312.1606*** |
| D.Trade | -23.7065*** | -23.5561*** | 313.5471*** | -19.8242*** | -17.9817*** | 244.3373*** |
| Urban | -2.1992 ** | 9.3489 | 412.7868*** | -0.2494 | 1.3746 | 385.7663*** |
| D.Urban | -8.9989*** | -5.8673*** | 391.6831*** | -25.6282*** | -10.3542*** | 516.4218*** |
| Lnpop | -8.6576 *** | 0.5530 | 217.7913* | -7.9519*** | -9.5615*** | 171.6226 |
| D.Lnpop | -4.8050*** | -5.9777*** | 263.3420*** | -12.4152*** | -10.8649*** | 270.0260*** |
| Lnrd | -9.1802*** | -0.6381 | 252.3247*** | -46.4066*** | -13.6065*** | 184.2008 |
| D.Lnrd | -52.2230*** | -33.7355*** | 321.3852*** | -46.9865*** | -28.4670*** | 266.8772*** |
| Indu | -5.6185 *** | -1.8981** | 283.3107*** | -68.2961*** | -10.2091*** | 250.0945*** |
| D.Indu | -76.8840*** | -32.6631*** | 401.3513*** | -8.1e + 02*** | -1.0e + 02*** | 365.9560*** |
4.2 Panel cointegration test
After determining the panel stationarity of each series, we then performed three different tests, namely, the Kao, Pedroni and Westerlund tests (Kao 1999, Pedroni 1999, Westerlund 2007), to detect the presence of cointegration. The general idea is to test the null hypothesis by inferring whether the error correction term in the conditional error correction model is equal to zero. Rejecting the null hypothesis without error correction means rejecting the null hypothesis without cointegration. When T (time) < N (number of cross sections), as in this study, the tests have good statistical properties relative to other tests and can be used with balanced and unbalanced panels.
As shown in Table 3, the no cointegration null hypothesis can be rejected for all three major estimation algorithms and for all the sub-Kao tests and sub-Pedroni tests. The conclusion of rejecting the null hypotheses remains consistent even for models with and without the proxy for climate finance components (mitigation climate finance and adaptation climate finance). It can therefore be concluded that the results exhibit robust support for a long-term linkage between the variables.
| Test Type | Intercept | Intercept & Trend | |
|---|---|---|---|
| Kao test | Modified Dickey-Fuller t | 2.0249** | -- |
| Dickey-Fuller t | 2.7226*** | -- | |
| Augmented Dickey-Fuller t | 4.2442*** | -- | |
| Pedroni test | Modified Phillips-Perron t | 13.0633*** | 15.1388*** |
| Phillips-Perron t | -12.5633*** | -14.8343*** | |
| Augmented Dickey-Fuller t | -12.2651*** | -14.9112*** | |
| Westerlund test | -2.6638*** | 0.3829 | |
5.1 Descriptive statistics
Table 4 reports summary statistics on all variables covering the mean, minimum, maximum and standard deviation. The mean log value of carbon emissions intensity is 6.116, indicating 610.81 ton/$million, which is approximately 22% higher than the global average. The Std. Dev. of Lntcf is much higher than that of Lnce, indicating that the climate aid of 93 BRI countries is unbalanced, and there are great differences among regions. To determine whether multicollinearity is present in the data, we evaluated the variance inflation factor (VIF). (Kennedy 1992) argued that when the VIF is less than 10, regressors in the model are free from multicollinearity. In our data, all variables meet these criteria, indicating that multicollinearity is not an issue.
|
Variable |
Mean |
Min |
Max |
Std. Dev. |
VIF |
|---|---|---|---|---|---|
|
Lnce |
6.116 |
3.943 |
8.520 |
0.745 |
-- |
|
Lntcf |
8.318 |
0 |
15.165 |
4.245 |
1.32 |
|
Lnpgdp |
7.886 |
5.351 |
10.195 |
1.086 |
3.31 |
|
Lnpop |
15.876 |
11.151 |
21.055 |
1.865 |
3.86 |
|
Lnrd |
5.031 |
-20.723 |
13.177 |
2.98 |
2.81 |
|
Trade |
79.546 |
0.17 |
311.35 |
36.469 |
1.44 |
|
Urban |
49.619 |
8.25 |
95.33 |
20.04 |
2.06 |
|
Indu |
0.28 |
0.032 |
0.878 |
0.135 |
1.70 |
5.2 Benchmark regression
The panel data of 93 selected BRI countries were used for benchmark regression, and the results are shown in Table 5. Columns 1 and 2 show the results of the OLS model and the fixed effects model, respectively, and Columns 3 and 4 show the results of two-step difference GMM and two-step system GMM, respectively. Although the coefficients of the lagged dependent variable by difference GMM and system GMM are both statistically significant, the coefficient of system GMM is between the coefficient of OLS and fixed effects, while the coefficient of difference GMM is not, which indicates that the dynamic model is applicable to system GMM. Therefore, further analysis will illustrate the results of system GMM estimator.
|
(1) |
(2) |
(3) |
(4) |
(5) |
(6) |
|
|---|---|---|---|---|---|---|
|
Variable |
OLS |
LFE |
DifGMM |
SysGMM |
SysGMM |
SysGMM |
|
L.Lnce |
0.975*** |
0.809*** |
0.531*** |
0.896*** |
0.898*** |
0.904*** |
|
(211.24) |
(50.79) |
(4.82) |
(28.05) |
(27.57) |
(29.03) |
|
|
Lntcf |
-0.000 |
-0.001 |
-0.002 |
-0.004*** |
||
|
(-1.53) |
(-2.77) |
|||||
|
Lnmtcf |
-0.003** |
|||||
|
(-2.51) |
||||||
|
Lnadcf |
-0.003* |
|||||
|
(-1.84) |
||||||
|
Lnpgdp |
0.063 |
0.720*** |
0.270 |
0.280*** |
0.268*** |
0.237** |
|
(1.55) |
(5.18) |
(0.63) |
(2.95) |
(2.84) |
(2.57) |
|
|
Lnpgdp2 |
-0.005* |
-0.049*** |
-0.050** |
-0.018*** |
-0.017*** |
-0.015** |
|
(-1.75) |
(-5.47) |
(-2.02) |
(-2.92) |
(-2.82) |
(-2.54) |
|
|
Lnpop |
0.001 |
0.064 |
-0.058 |
0.014** |
0.014** |
0.012* |
|
(0.18) |
(1.29) |
(-0.37) |
(2.15) |
(2.10) |
(1.94) |
|
|
Lnrd |
-0.001 |
-0.003 |
-0.004*** |
-0.002 |
-0.002 |
-0.002 |
|
(-0.51) |
(-1.08) |
(-4.00) |
(-0.50) |
(-0.50) |
(-0.71) |
|
|
Trade |
0.000** |
0.000* |
0.002*** |
0.001 |
0.001 |
0.001* |
|
(2.58) |
(1.81) |
(2.83) |
(1.53) |
(1.53) |
(1.69) |
|
|
Urban |
0.000 |
0.002 |
0.003 |
0.001 |
0.001 |
0.000 |
|
(0.00) |
(0.99) |
(0.90) |
(1.35) |
(1.34) |
(1.25) |
|
|
Indu |
0.054* |
-0.027 |
-0.192 |
0.066 |
0.067 |
0.077 |
|
(1.95) |
(-0.34) |
(-1.13) |
(1.14) |
(1.18) |
(1.29) |
|
|
AR1(P) |
0.000 |
0.000 |
0.000 |
0.000 |
||
|
AR2(P) |
0.563 |
0.444 |
0.450 |
0.458 |
||
|
Hansen(P) |
0.727 |
0.789 |
0.776 |
0.121 |
||
|
R-squared |
0.976 |
0.979 |
||||
|
Number of id |
93 |
93 |
93 |
93 |
93 |
93 |
|
Observations |
1,674 |
1,674 |
1,581 |
1,674 |
1,674 |
1,674 |
| Source: Author estimations. | ||||||
| Note: ***, **, * represent levels of significance at 1%, 5% and 10%, respectively. Robust t-statistics are reported in parentheses of estimated coefficients. GMM-type variables are Lnce and Trade. Lags' range of Lnce is set from one to three, and Trade is set to two lags of itself in all models. | ||||||
The results of system GMM in Column 4 show that the estimated coefficient on total climate finance is negative and statistically significant at 1%. Thus, a percentage increase in total climate finance decreases the CO2 emissions intensity in BRI countries by 0.004%. Using the total climate finance, GDP and carbon emissions data of 93 BRI countries from 2000 to 2018, it is further estimated that every one billion USD of climate finance can reduce carbon emissions intensity by 0.016 ton/million USD. The non-linear effect of PGDP shows that the estimated coefficient is statistically significant at 1%. The main term of economic growth exerts a significant positive impact on carbon emissions intensity, while its squared term exerts a significant negative effect, indicating the existence of the EKC hypothesis and suggesting that carbon emissions intensity will first increase and then decrease with economic growth in BRI countries. This result is also in line with previous empirical studies of the relationship between economic growth increases and carbon emissions (Bento and Moutinho 2016, Liobikienė and Butkus 2019). The regression coefficient of population is positive at the 5% significance level, indicating that population agglomeration increases the level of carbon emissions. The regression coefficients of the other explanatory variables RD, TRADE, URBAN and INDU are not significant.
Columns 5 and 6 of Table 5 show the impacts of the two components of climate finance, mitigation climate finance and adaptation climate finance, on the carbon emissions of BRI countries. The estimated coefficient of mitigation climate finance is negative and statistically significant at 5%. Thus, a percentage increase in mitigation climate finance decreases the carbon emissions intensity of BRI countries by 0.003%, which is slightly lower than the coefficient of total climate finance. Mitigation climate finance is approximately $229.1 billion, with more than 40,000 support projects, accounting for 71.1% of total climate finance. For the supporting areas, energy, transport and storage, and agriculture, forestry and fishing with carbon emissions reduction benefits are priorities, with $90.3 billion, $42.9 billion, and $10.4 billion in capital scale and 7,660, 1,060, and 6,571 in project scale, respectively. Therefore, mitigation is not only the main usage of climate finance but also the main source of carbon emissions reduction. The regression results show that the estimated coefficient of adaptation climate finance is also negative and statistically significant. Thus, adaptation climate finance also has reduction effects on the carbon emissions intensity of BRI countries. Adaptation climate finance is approximately $121.6 billion with approximately 24,000 support projects, much less than mitigation climate finance. As the supporting areas mainly focus on water supply and sanitation, agriculture, forestry, fishing, and disaster prevention and preparedness that are not primarily aimed at reducing GHG emissions, the carbon emissions reduction effects are not as significant as mitigation climate finance.
5.4 Heterogeneity test
To further study the impact of climate aid on carbon emissions, this section focuses on analysing regional heterogeneity. 93 BRI countries are divided into three regions: Africa (afri), Asia (asia), and Other (othe, including Europe, Oceania and America) (see Appendix A). Following the heterogeneity analysis by (Yu and Zhang 2021), three dummy variables (afri, asia, and othe) are introduced into the baseline model to capture the heterogeneity impacts of climate finance. The variable afri equals one if a country is in Africa; otherwise, it is equal to zero. It is similar for the variable asia and othe. Table 6 shows the results of our regression. Our findings suggest that both the interaction term in Africa and in Europe, Oceania and America, have significantly negative coefficients, while the results of Asia is not significant. The total amount of climate finance in Asia is $149.44 billion, with more than 20,814 support projects, accounting for nearly half of total climate finance (46.4%). Besides, the climate finance flows into areas with emission reduction effects, such as energy, storage and transport are more than the average of BRI region. However, many Asian countries are in the process of rapid economic growth with high level carbon emissions, which offsets the emission reduction effects that climate finance brings. Therefore, policymakers in Asian countries should strengthen and promote the low-carbon transition as soon as possible. On the one hand, more efforts should be made for the low-carbon transitions in industry, transport, residential sector and etc. On the other hand, energy structure should be gradually transformed from fossil fuels to renewable energy and energy efficiency should be improved. It also shows that the climate finance in Africa has greater impact on carbon emissions reduction.
|
Variable |
Lnce |
|---|---|
|
L.lncei |
0.895*** |
|
(29.42) |
|
|
lntcf×afri |
-0.006*** |
|
(-3.03) |
|
|
lntcf×asia |
-0.001 |
|
(-0.69) |
|
|
lntcf×othe |
-0.005*** |
|
(-2.93) |
|
|
Constant |
-0.987** |
|
(-2.62) |
|
|
Control variables |
Y |
|
Fixed effects |
Y |
|
AR1(P) |
0.000 |
|
AR2(P) |
0.425 |
|
Hansen(P) |
0.812 |
|
Number of id |
93 |
|
Observations |
1,674 |
| Source: Author estimations. | |
| Note: ***, **, * represent levels of significance at 1%, 5% and 10%, respectively. Robust t-statistics are reported in parentheses of estimated coefficients. GMM-type variables are Lnce and Trade. Lags' range of Lnce is set from one to three, and Trade is set to two lags of itself in all models. | |
5.5 Mechanism test
To further test the impact mechanism of climate aid on carbon emissions, this paper selects the percentage of renewable energy production to total energy production (RE) as an intermediary variable. The empirical results of the mediating effect test for both total climate finance (TCF) and mitigation climate finance (MTCF) are shown in Table 7. Columns 1–2 and 3–4 report the mediating effect of TCF and MTCF, respectively. First, as we analysed in Section 5.1, the coefficients of Lntcf and Lnmtcf are − 0.004 and − 0.003, respectively, and significantly negative at the 1% level and 5% level, respectively. Second, the results in Columns 1 and 3 demonstrate that the coefficients of Lntcf and Lnmtcf on RE are both 0.006 and significant at the 1% level, which means that an increase in climate aid will lead to an increase in the proportion of renewable energy in BRI countries. Third, the results presented in Columns 2 and 4 show that the coefficients of RE to Lnce are both − 0.003 and significantly negative at the 1% level, which means that an increase in the share of renewable energy will significantly reduce carbon emissions intensity in BRI countries, and this conclusion is consistent with most existing studies (Khan et al. 2020, Yu et al. 2018). It can then be concluded that climate aid could help BRI countries achieve carbon emissions reduction through the mediating effect of increasing the proportion of renewable energy.
|
TCF |
MTCF |
|||
|---|---|---|---|---|
|
(1) |
(2) |
(3) |
(4) |
|
|
Variable |
RE |
Lnce |
RE |
Lnce |
|
L.Lnce |
0.900*** |
0.900*** |
||
|
(30.13) |
(30.55) |
|||
|
L.RE |
0.612*** |
0.612*** |
||
|
(9.16) |
(9.14) |
|||
|
RE |
-0.040** |
-0.040** |
||
|
(-2.35) |
(-2.35) |
|||
|
Lntcf |
0.006*** |
-0.003** |
||
|
(3.29) |
(-2.42) |
|||
|
Lnmtcf |
0.006*** |
-0.003** |
||
|
(3.44) |
(-2.07) |
|||
|
Control variables |
Y |
Y |
Y |
Y |
|
Fixed effects |
Y |
Y |
Y |
Y |
|
AR1(P) |
0.014 |
0.000 |
0.014 |
0.000 |
|
AR2(P) |
0.143 |
0.430 |
0.146 |
0.442 |
|
Hansen(P) |
0.874 |
0.685 |
0.879 |
0.663 |
|
Number of id |
93 |
93 |
93 |
93 |
|
Observations |
1,674 |
1,674 |
1,674 |
1,674 |
| Source: Author estimations. | ||||
| Note: ***, **, * represent levels of significance at 1%, 5% and 10%, respectively. Robust t-statistics are reported in parentheses of estimated coefficients. GMM-type variables are Lnce and Trade. Lags' range of Lnce is set from one to three, and Trade is set to two lags of itself in all models. | ||||
The total carbon emissions reduction effect can be divided into direct effects and indirect effects, and the results are shown in Table 8 and Table 9. For TCF, the direct effects c’=-0.020 and the indirect effects ab=-0.005. For MTCF, the direct effects c’=-0.016 and the indirect effects ab=-0.006. c’ has the same sign as ab, indicating that renewable energy does play a mediating effect on the relationship between climate aid and carbon emissions in BRI countries. The indirect effects account for 21.99% and 29.35% of the total effects in the two cases where Lntcf and Lnmtcf are taken as dependent variables, respectively, while the ratios of absolute indirect effects to direct effects are 0.28 and 0.41 in the Lntcf and Lnmtcf cases, respectively. This result means that the indirect effect is smaller than the direct effect and that the indirect effect in the Lnmtcf case is more significant than that in the Lntcf case. There are two reasons for this finding. One is that the climate aid used for energy was $90.3 billion, accounting for 28% and 39% of the total climate finance and mitigation climate finance, respectively, indicating that energy projects partially contribute to carbon emissions reduction. The other is that the indirect effects achieved by renewable energy require relatively advanced low-carbon technology and a foundation for clean energy development in recipient countries, which take some time to realize.
|
Indicator |
TCF |
MTCF |
|---|---|---|
|
Sobel test |
-0.005*** (-5.013) |
-0.007*** (-5.537) |
|
Aroian test |
-0.005*** (-4.993) |
-0.007*** (-5.517) |
|
Goodman test |
-0.005*** (-5.033) |
-0.007*** (-5.557) |
|
Indirect effect |
-0.005*** (-5.013) |
-0.007*** (-5.537) |
|
Direct effect |
-0.020*** (-4.713) |
-0.016*** (-3.68) |
|
Total effect |
-0.025*** (-5.95) |
-0.023*** (-5.146) |
|
Proportion of total effect that is mediated |
0.22 |
0.29 |
|
|Indirect/Direct| |
0.28 |
0.41 |
| Source: Author estimations. | ||
| Note: ***, **, * represent levels of significance at 1%, 5% and 10%, respectively. Robust t-statistics are reported in parentheses of estimated coefficients. | ||
|
Independent variable |
Effects |
Coef. |
Std. Err. |
95% conf. interval |
|
|
lower value |
upper value |
||||
|
TCF |
Indirect |
-0.005 |
0.001 |
-0.007 |
-0.004 |
|
Direct |
-0.020 |
0.005 |
-0.030 |
-0.009 |
|
|
MTCF |
Indirect |
-0.007 |
0.001 |
-0.009 |
-0.005 |
|
Direct |
-0.016 |
0.006 |
-0.027 |
-0.005 |
|
| Source: Author estimations. | |||||
5.3 Robustness test
In the benchmark regression, the results of the OLS and fixed effect methods are significant except for the GMM estimator we selected, preliminarily indicating the robustness of the empirical results. Here, we further corroborate the robustness in two ways. One is to narrow the time scale and examine whether the results are statistically significant. We noticed that in the 2009 Copenhagen Climate Change Conference, developed countries pledged to raise $100 billion per year for developing countries by 2020. After that, the scale of climate aid significantly increased from $7.29 billion in 2009 to $40.32 billion in 2018 (see Fig. 2 in the Introduction section). Therefore, this paper selected data from the period 2009–2018 for the robustness test, and the results are shown in Column 1 of Table 10. It is estimated that the coefficient of Lntcf is -0.005 and statistically significant at 1%, of which the absolute value is slightly higher than the benchmark regression (-0.004). The reason for this finding is that the annual climate finance during the period of 2009–2018 is higher than that during the period of 2000–2018. Therefore, the higher the climate finance invested, the better the carbon emissions reduction that can be achieved.
The other method of robustness testing is to use the current price of climate finance instead of a constant price. Climate finance data sourced from the CRS database of OECD-DAC are described in both constant price and current price. As the value in constant price is used for the benchmark regression, it is necessary to use the value in current price for the robustness test. The results of the robustness test are shown in Column 2 of Table 10. It is estimated that the coefficient of Lntcf’ is -0.004 and statistically significant at 1%, of which the absolute value is the same as the benchmark regression. The value of climate finance at a constant price is not much different from that at the current price, so the coefficient is the same. According to the robustness test results, it can be concluded that the positive effects of climate aid on carbon emissions reduction in BRI countries are robust and reliable.
|
(1) |
(2) |
|
|---|---|---|
|
SysGMM1 |
SysGMM2 |
|
|
Variable |
Lnce |
Lnce |
|
L.Lnce |
0.899*** |
0.896*** |
|
(30.62) |
(27.93) |
|
|
Lntcf |
-0.005** |
|
|
(-2.06) |
||
|
Lntcf’ |
-0.004*** |
|
|
(-2.72) |
||
|
Constant |
-1.003* |
-1.194*** |
|
(-1.72) |
(-2.82) |
|
|
Control variables |
Y |
Y |
|
Fixed effects |
Y |
Y |
|
AR1(P) |
0.000 |
0.000 |
|
AR2(P) |
0.939 |
0.442 |
|
Hansen(P) |
0.424 |
0.787 |
|
Number of id |
93 |
93 |
|
Observations |
930 |
1,674 |
| Source: Author estimations. | ||
| Note: ***, **, * represent levels of significance at 1%, 5% and 10%, respectively. Robust t-statistics are reported in parentheses of estimated coefficients. GMM-type variables are Lnce and Trade. Lags' range of Lnce is set from one to three, and Trade is set to two lags of itself in all models. | ||
6.1 Conclusions
To achieve the global carbon neutrality target, the low-carbon transition of BRI countries is of great importance and requires a large amount of financial support for deploying extensive climate-related projects. BRI countries have received over 60% of global climate aid, but few studies have evaluated the impact of climate aid on carbon emissions and the mechanism involved. Therefore, this study first established a systematic panel approach using the two-step system GMM estimator and then used the panel data of 93 BRI countries from 2000–2018 for empirical analysis.
The empirical results show that climate aid has significant positive effects on carbon emissions reduction in BRI countries, with a percentage increase in total climate finance decreasing the CO2 emissions intensity by 0.004%. It is further estimated that every one billion USD invested in climate finance can reduce carbon emissions intensity by 0.016 ton/$million. When the panel data during the period of 2009–2018 is used for empirical analysis, it shows a slightly higher decreasing of the CO2 emissions intensity by 0.005%. The reason is that the annual climate finance during the period of 2009–2018 is higher than that during the period of 2000–2018. Therefore, the higher the climate finance invested, the better the carbon emissions reduction that can be achieved.
The empirical results of the two components of climate finance, mitigation climate finance and adaptation climate finance, show that both have positive effects on carbon emissions reduction in BRI countries. A percentage increase in mitigation climate finance decreases the carbon emissions intensity by 0.003%, which is slightly lower than the coefficient of total climate finance. The mitigation finance accounts for 71.1% of the total climate finance and mainly supports energy, transport and storage, and agriculture, forestry and fishing with carbon emissions reduction benefits. Therefore, mitigation is not only the main usage of climate finance but also the main source of carbon emissions reduction. In addition, adaptation climate finance also has reduction effects on the carbon emissions intensity of BRI countries. As adaptation climate finance mainly support areas that are not primarily aimed at reducing GHG emissions, the carbon emissions reduction effects are not as significant as mitigation climate finance.
The impact mechanism analysis indicates that climate aid has dual carbon emissions reduction effects in BRI countries with an intermediate energy structure. Climate aid could help BRI countries achieve carbon emissions reduction through the mediating effect of increasing the proportion of renewable energy. This means that climate aid not only directly reduces carbon emissions by increasing carbon reduction resources in BRI countries but also indirectly reduces carbon emissions by promoting renewable energy to optimize the energy structure. However, the indirect effect is smaller than the direct effect, as the indirect carbon reduction effect achieved by renewable energy requires relatively advanced low-carbon technology and a foundation for clean energy development in recipient countries, which take some time to realize.
6.2 Policy implications
First, developed countries should continue to increase the scale of climate finance support to developing countries along the Belt and Road. The recently released climate agreement from the Glasgow Climate Convention strengthens the climate finance target that developed countries not only meet their $100 billion climate finance commitments but also provide a cumulative $500 billion over the coming five-year period until 2025. The U.S.–China Joint Glasgow Declaration on Enhancing Climate Action in the 2020s also recognizes the importance of the commitment made by developed countries to the goal of jointly mobilizing $100 billion per year by 2020 and annually through 2025 to address the needs of developing countries and stresses the importance of meeting that goal as soon as possible. Therefore, developed countries should do as the commitment to provide more climate finance support as soon as possible.
Second, the effective usage of climate aid should be improved from two aspects. On the one hand, the additionality and the expected emissions reduction effects should be carefully evaluated before investment. The most remarkable characteristic of climate projects is additionality. The projects that are intended for construction regardless of whether there is climate aid should not be in the supporting scope of climate aid. Therefore, it is necessary to evaluate additionality in addition to climate benefits. On the other hand, as carbon emissions reduction is the core of achieving carbon neutrality and energy consumption is the main driver of carbon emissions, investment for energy should be further strengthened in areas such as renewable energy, energy efficiency, and fossil fuel transition. To leverage more investment from the private sector, in addition to public funding from developed countries, should also be encouraged.
Third, data monitoring and performance evaluation of climate aid should be strengthened. Climate aid tracking methods should be improved to accurately count the amount of finance flow into climate change, which involves avoiding double counting and finance attribution. A database of climate aid should be built for BRI countries to reflect the climate aid received from the vertical time scale and the horizontal space scale. Performance evaluation indicators of climate aid funds, including mitigation indicators such as the carbon emissions reduction effect and adaptation indicators such as the number of beneficiaries of climate adaptation, should be designed to accurately evaluate the effect of climate aid. An adjustment mechanism should be established to dynamically adjust the scale and flow of climate aid according to the climate change goals and the progress of BRI countries.
6.3 Limitations and future research
Although this paper enriches the study of the impact and mechanism of climate aid on carbon emissions, some limitations might deserve further investigation. First, due to data availability, only climate aid, not total financial support, is taken into consideration. In fact, climate finance also comes from multilateral sources, private sectors and domestic finance. Further research considering more finance data should be conducted to better evaluate the effects of climate-related finance. Second, the share of renewable energy production is selected as a mediating variable in this study, and indicators such as technological innovation and industrial transformation can also be considered to explore the mediating effect. Finally, the emissions reduction effects of climate finance can be further estimated with quantitative methods to better evaluate the performance of climate finance and provide guidance for carbon neutrality strategies.
Acknowledgements: The authors appreciate the editor and reviewers for useful comments and suggestions.
Author contribution An Zeng: conceptualization, data collection and analysis, manuscript writing. Yuhui Sheng: data collection, econometric model and empirical analysis.Baihe Gu: conceptualization, results analysis, manuscript review and editing.Zhengzao Wang: econometric model and empirical analysis. Mingyue Wang: econometric model and empirical analysis.
Funding: This research was funded by the National Key Research and Development Program of China (2018YFA0606504) and National Natural Science Foundation of China (72140007,71804178).
Data availability: Data are available on request.
Ethics approval: The study did not use any data which need approval.
Consent to participate: All authors participated in the process of draft completion. All authors have read and agreed to the published version of the manuscript.
Consent to publish: All authors agree to publish.
Competing interests: The authors declare no competing interests.
- Arellano M, Bond S (1991): Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. Rev. Econ. Stud. 58, 277–297
- Bento JPC, Moutinho V (2016): CO2 emissions, non-renewable and renewable electricity production, economic growth, and international trade in Italy. Renew. Sustain. Energy Rev. 55, 142-155
- Bhattacharyya S, Intartaglia M, McKay A (2018): Does energy-related aid affect emissions? Evidence from a global dataset. Review of Development Economics 22, 1166-1194
- Blundell R, Bond S (1995): Initial Conditions and Moment Restrictions in Dynamic Panel Data Models, Economics Papers 104, Economics Group, Nuffield College, University of Oxford
- Blundell R, Bond S (1998): Initial Conditions and Moment Restrictions in Dynamic Panel Data Models. Journal of Econometrics 87, 115-143
- Bosah CP, Li S, Ampofo GKM, Liu K (2021): Dynamic nexus between energy consumption, economic growth, and urbanization with carbon emission: evidence from panel PMG-ARDL estimation. Environ Sci Pollut Res Int 28, 61201-61212
- Carfora A, Ronghi M, Scandurra G (2017): The effect of Climate Finance on Greenhouse Gas Emission: A Quantile Regression Approach. International Journal of Energy Economics and Policy 7, 185-199
- Carfora A, Scandurra G (2019): The impact of climate funds on economic growth and their role in substituting fossil energy sources. Energy Policy 129, 182-192
- Charlene Watson LS (2021): The Global Climate Architecture 2020.
- Choi I (2001): Unit root tests for panel data. Journal of International Money and Finance 20, 249-272
- Chung YW, Shin KN, Sohn SH (2018): The effects of climate technology-related development finance on greenhouse gases reduction. Korea and the World Economy 19, 105-135
- Copeland BR, Taylor MS (1994): North-south trade and the environment. Quarterly Journal of Economics 109, 755-787
- EIA (2021): RENEWABLE & ALTERNATIVE FUELS
- Ellis J, Caruso R, Ockenden S (2013): Exploring Climate Finance Effectiveness., OECD/IEA Climate Change Expert Group Papers (No. 2013/04). OECD Publishing, Paris
- Fan J-L, Da Y-B, Wan S-L, Zhang M, Cao Z, Wang Y, Zhang X (2019): Determinants of carbon emissions in ‘Belt and Road initiative’ countries: A production technology perspective. Applied Energy 239, 268-279
- Gu A, Zhou X (2020): Emission reduction effects of the green energy investment projects of China in belt and road initiative countries. Ecosystem Health and Sustainability 6, 1747947
- Han M, Lao J, Yao Q, Zhang B, Meng J (2020): Carbon inequality and economic development across the Belt and Road regions. Journal of Environmental Management 262, 110250
- Hu Y, Ren S, Wang Y, Chen X (2020): Can carbon emission trading scheme achieve energy conservation and emission reduction? Evidence from the industrial sector in China. Energy Economics 85
- Hussain MN, Li Z, Sattar A (2021): Effects of urbanization and nonrenewable energy on carbon emission in Africa. Environ Sci Pollut Res Int
- Im KS, Pesaran MH, Shin Y (2003): Testing for unit roots in heterogeneous panels. Journal of Econometrics 115, 53-74
- IPCC (2021): Summary for Policymakers. In: Masson-Delmotte V, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J.B.R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu and B. Zhou (eds.) (Editor), Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge
- IRENA 2018: Global energy transformation: a road map to 2050
- Kao C (1999): Spurious Regression and Residual-Based Tests for Cointegration in Panel Data. Journal of Econometrics 90, 1-44
- Kennedy P (1992): A Guide to Econometrics, Blackwell, Oxford
- Khan I, Han L, Khan H (2021): Renewable energy consumption and local environmental effects for economic growth and carbon emission: evidence from global income countries. Environ Sci Pollut Res Int
- Khan Z, Ali S, Umar M, Kirikkaleli D, Jiao Z (2020): Consumption-based carbon emissions and International trade in G7 countries: The role of Environmental innovation and Renewable energy. Science of The Total Environment 730, 138945
- Kretschmer B, Hübler M, Nunnenkamp P (2013): Does foreign aid reduce energy and carbon intensities of developing economies? Journal of International Development 25, 67-91
- Levin A, Lin C-F, Chu C-SJ (2002): Unit root tests in panel data: Asymptotic and finite sample properties. Journal of Econometrics 108, 1–24
- Lin B, Ahmad I (2017): Analysis of energy related carbon dioxide emission and reduction potential in Pakistan. Journal of Cleaner Production 143, 278-287
- Liobikienė G, Butkus M (2019): Scale, composition, and technique effects through which the economic growth, foreign direct investment, urbanization, and trade affect greenhouse gas emissions. Renewable Energy 132, 1310-1322
- Munoz C, Miquel KG, Zhongshu L (2018): Renewable Energy: The Trillion Dollar Opportunity for Chinese Overseas Investment. China and World Economy 26, 27-49
- Naito T (2003): Pareto-improving untied aid with environmental externalities. Journal of Economics 80, 161-169
- OECD (2021): Climate Change: OECD DAC External Development Finance Statistics
- Omri A, Bel Hadj T (2020): Foreign investment and air pollution: Do good governance and technological innovation matter? Environ Res 185, 109469
- Ono T (1998): Consumption externalities and the effects of international income transfers on the global environment. Journal of Economics 68, 255-269
- OWD (2021): CO₂ and Greenhouse Gas Emissions
- Pedroni P (1999): Critical Values for Cointegration Tests in Heterogeneous Panels with Multiple Regressors. Oxford Bulletin of Economics and Statistics 61, 653-670
- Pei C, Wang X, Lin Y, Fang L, Meng S (2019): Inhibition of Galectin-3 Alleviates Cigarette Smoke Extract-Induced Autophagy and Dysfunction in Endothelial Progenitor Cells. Oxidative Medicine and Cellular Longevity 2019, 7252943
- Qu X, Luo H (2021): Impact of China’s OFDI on carbon emissions and its transmission mechanism: an empirical analysis based on multiple mediation effect model. China population，resources and environment 31, 1-14
- She Q, Wu X (2019): Effects of Energy-Related Aid on Carbon Emissions in Recipient Countries: An Empirical Analysis of 26 Countries in Asia. Ecological Economy 35
- Shen Y, Su ZW, Malik MY, Umar M, Khan Z, Khan M (2021): Does green investment, financial development and natural resources rent limit carbon emissions? A provincial panel analysis of China. Sci Total Environ 755, 142538
- Tao Y, Liang H, Celia MA (2020): Electric power development associated with the Belt and Road Initiative and its carbon emissions implications. Applied Energy 267, 114784
- UN (1997): Kyoto Protocal.
- Wang W, She Q (2018): Climate Aid and the Carbon Dioxide Emissions of Recipient Countries: The Theoretical Mechanism and Empirical Research Based on Chinese Provincial Panel Data. Ecological Economy 34, 21-27
- Wang Z, Zhu Y, Zhu Y, Shi Y (2016): Energy structure change and carbon emission trends in China. Energy 115, 369-377
- WDI (2021): World Bank Open Data
- Wen Z, Ye B (2014): Analyses of mediating effects: the development of methods and models. Adv. Psychol. Sci. 22, 731-745
- Westerlund J (2007): Testing for error correction in panel data. Oxf. Bull. Econ. Stat. 69, 709–748
- Wu J, Tang L, Mohamed R, Zhu Q, Wang Z (2015): Modeling and assessing international climate financing. Frontiers of Earth Science 10, 253-263
- Wu X, Pan A, She Q (2021): Direct and indirect effects of climate aid on carbon emissions in recipient countries. Journal of Cleaner Production 290
- Yang F, Yang M (2019): Greening the one belt and one road initiative. Mitigation and Adaptation Strategies for Global Change 24, 735-748
- Yu S, Zheng S, Li X (2018): The achievement of the carbon emissions peak in China: The role of energy consumption structure optimization. Energy Economics 74, 693-707
- Yu Y, Zhang N (2021): Low-carbon city pilot and carbon emission efficiency: Quasi-experimental evidence from China. Energy Economics 96
- Zhang X, Zhang H, Zhao C, Yuan J (2019): Carbon emission intensity of electricity generation in Belt and Road Initiative countries: a benchmarking analysis. Environmental Science and Pollution Research 26, 15057-15068
- Zhu C, Gao D (2019): A Research on the Factors Influencing Carbon Emission of Transportation Industry in “the Belt and Road Initiative” Countries Based on Panel Data. Energies 12