Synergistic policy construction of carbon mechanisms and carbon tariffs in the transition to a low-carbon economy

The increasing tension of global energy and environmental problems has given rise to the concept of low carbon economy, and the developmental economic system aiming at low energy consumption, low pollution and low emission has been recognized globally [1]. From the international trend, the development of low carbon economy and circular economy, realize green recovery, has become a world trend. At present, the international community has realized or has put forward the goal of carbon neutrality of 31 countries, is in the pipeline to put forward the goal of carbon neutrality of nearly hundreds of countries, to realize the trend towards carbon neutrality has become unquestionable [2]. Some countries and regions have begun to announce specific implementation plans one after another, such as the European Commission released the European Green Deal in December 2019, emphasizing the realization of carbon neutrality in 2050 as the core strategic objective, and constructing a modern competitive economic system that decouples economic growth from resource consumption [3]. In September 2020, China put forward the carbon peak and carbon neutrality targets, which set a green tone for the long-term development of China’s economy, and the Chinese government will ensure that the dual-carbon targets will be realized on schedule with the help of the “1+N” policy system, which will usher in a period of intensive release of green environmental policies [4-6]. China’s rapid development in recent years has undoubtedly made great contributions to the global economy. However, being on the cusp of environmental protection, the high-speed industrialization progress also brings certain energy and environmental drawbacks, and actively responding to the advocacy of low-carbon economy and transitioning to sustainable economic development has become an important turning point for China’s economic growth [7-9]. So far, globally, more and more countries and regions have begun to move towards low-carbon economic transformation to cope with the increasingly severe climate change and environmental pollution problems [10].

Carbon mechanisms and carbon tariffs, as policy tools in the transition to a low-carbon economy, cannot be ignored in reducing carbon emissions, promoting the development of a green economy and actively responding to climate change [11-12]. Carbon mechanism, specifically includes carbon market and carbon trading mechanism. In the carbon market, according to the demand for carbon emission reduction, the government divides a certain amount of carbon emission rights and allocates its interests to enterprises and organizations. Carbon emissions are regarded as a commodity, and enterprises can realize the flexibility on emissions by buying or selling carbon emission rights, so as to achieve the goal of reducing carbon emissions [13]. Enterprises can buy or sell carbon emission rights according to their own needs to achieve their own carbon emission reduction goals, and this carbon trading mechanism is usually based on market supply and demand without strict government intervention [14]. Instead, the government strictly monitors and verifies the carbon emissions of enterprises by setting carbon emission quotas and trading rules, and enterprises must purchase a sufficient number of carbon emission rights within a specified period of time or face fines or other penalties [15]. This has led to the implementation of the carbon trading market in a way that involves a variety of factors, but the most critical are the pricing mechanism of carbon emission rights, market trading mechanism and policy support [16]. For this reason, it also constrains the low-carbon economic transition.

In addition, from the surface analysis of protecting the environment and saving energy, the introduction of carbon tariffs has its reasonableness and positivity to reduce the emission of greenhouse gases, control the global warming trend, adjust the industrial structure, develop a low-carbon economy, and realize sustainable development [17-19]. But the United States at this time put forward the idea of carbon tariffs, but also has its other reality of deep meaning, one is by trying to improve the competitiveness of domestic products through the imposition of carbon tariffs to reduce the trade deficit [20]. At present, China and other emerging markets rapid economic development, relying on relatively low manufacturing costs in the export trade to occupy the price competition advantage. The United States and other developed countries hope that by proposing nearly demanding carbon emission standards, and accordingly impose tariffs, so that the United States in the field of energy conservation and environmental protection and new energy and other technologies between countries to occupy the high ground in the competition, to curb the rise of emerging countries such as China, Brazil, and other emerging countries, in particular, hope to use this to reverse the huge trade deficit with China over the years [21-23]. Secondly, it is trying to change its international image, so that it is in a favorable position in the global climate change negotiations [24]. The U.S. is a large energy consumer, and in terms of energy conservation and emission reduction the U.S. has always given people the impression of being rather negative and irresponsible. A reasonable carbon tariff policy is undoubtedly an important node in the transformation of low-carbon economy. As a result, the synergistic policy of exploring how the carbon mechanism and carbon tariffs are constructed in the context of low-carbon economic transformation is aimed at achieving the effect of energy conservation and emission reduction, promoting the green development of the country, and enhancing international competitiveness.

In the context of global response to climate change, low-carbon economic transformation has become an important direction of economic development in various countries. This paper chooses to construct the MF-VAR model as the data research model. It studies the key role of carbon mechanism and carbon tariffs as important policy tools to realize the emission reduction target and promote economic restructuring. The actual carbon emission data of industrial provinces are chosen as research examples. The MF-VAR model is built using trade intensity as an indicator of foreign trade, and per capita energy consumption and per capita carbon emissions as indicators of energy consumption and carbon emissions. Through the unit root test and other measurement steps, the variable data are gradually analyzed to see whether they have the characteristics of time series smoothness and so on. According to the research results, the dynamic relationship between low-carbon economic transformation and carbon mechanism and carbon tariff is revealed.

2

Analysis of mixed-frequency forecasting models for macroeconomic indicators

Accurate prediction of the dynamics of macroeconomic variables is crucial for policy formulation in the process of low-carbon economic transition. Although the traditional vector autoregression (VAR) model can effectively capture the dynamic relationships among variables, it is limited by the use of single frequency data. For this reason, this paper introduces the mixed-frequency vector autoregression (MF-VAR) model, which combines macroeconomic data of different frequencies to more accurately reflect the dynamic characteristics of the economic system. This part will introduce in detail the theoretical basis and construction ideas of VAR model and MF-VAR model, and analyze the advantages of the application of MF-VAR model in macroeconomic forecasting.

2.1

Vector autoregressive model

The vector autoregressive (VAR) model is a commonly used econometric model, which, along with its many extensions, is widely used in macroeconomic forecasting in various countries and regions and the world at large. The VAR model is a flexible time-series model that captures the complex dynamics of the relationships between macroeconomic variables, and it is also often used in macroeconomic forecasting. The VAR model is a model in which all of the current variables in the model are used to regression on a number of lagged variables of all variables, avoiding the need for structural modeling by treating each endogenous variable in the system as a function of the p-lagged values of all endogenous variables in the system. Among other things, the model allows for the use of only a single frequency of indicators, which are usually estimated based exclusively on quarterly observations or exclusively on monthly observations. The use of quarterly observations has the advantage of a larger macroeconomic series that may be included in the VAR model; the use of monthly observations has the advantage that the VAR model is able to track the economy more closely in real time. The VAR model is used to estimate the dynamics of joint endogenous variable relationships without any ex ante constraints. The model with n variables $V A R (p)$ $VAR\left( p \right)$ can be expressed as 1 $x_{t} = Φ_{1} x_{t - 1} + \dots + Φ_{p} x_{t - p} + Φ_{c} + u_{t}, u_{t} ~ i i d N (0, Σ)$ ${x_t} = {\Phi_1}{x_{t - 1}} + \cdots + {\Phi_p}{x_{t - p}} + {\Phi_c} + {u_t},{u_t}\sim iidN(0,\Sigma )$

where x_t is the n × 1 endogenous vector, x_t−1, ⋯, x_t−p is the n × 1 exogenous shock vector p is the lag order, Φ_c is the n × 1 constant coefficient column vector, Φ_i, i = 1, ⋯, p is the n × 1 coefficient column vector, u_t is the n × 1 error column vector, and the error terms are independently and identically distributed in a normal distribution with a mean of 0 and a variance of $\sum$ $\sum$ 0, and $\sum$ $\sum$ is a n×n-dimensional positive definite matrix. Assuming that the random error vectors are uncorrelated with the lag matrix and exogenous regressions, the standard orthogonality condition holds, and the posterior distribution of the coefficient matrices is centered on the ordinary least squares (OLS) estimates of the coefficients through the conditioning of the flattened a priori and the initial p observations, and is easy to compute. Furthermore, even if the random error vectors are correlated, all equations in the system have the same regression coefficients, so OLS is equivalent to generalized least squares (GLS) and the estimates are valid.

2.2

Mixed-frequency vector autoregressive model

Since traditional VAR models only allow the use of single frequency metrics, and considering the mismatch of sampling frequencies has become a standard for real-time forecasting versus short-term forecasting, in order to take advantage of the respective strengths of the annual VAR model and the quinquennial VAR model, this paper constructs a constant coefficients MF-VAR model. The MF-VAR model is an extended model of the VAR model, which allows for the different frequency of sampling, the MF-VAR model considered in this paper is based on the standard constant parameter VAR model described above, where variables are measured at annual frequencies. Since some macroeconomic time series are measured only at quinquennial frequencies, the corresponding annual series are treated as unobserved values in this paper. The low-frequency data are regarded as high-frequency data with intervals of missing, and a state-space model is built with the low-frequency quinquennial data and the high-frequency annual data that can be actually observed in each period as the observational variables, and with the potential annual data of each variable as the state variables, and the MF-VAR model is taken as the state equation of the state-space model, and the correspondence between the state variables and the observational variables is taken as the measurement equation of the state-space model.

Assuming that the economy develops at an annual frequency in equation (1), the vector of macroeconomic variables can be composed as $x_{t} = {[x_{m, t}^{'}, x_{q, t}^{'}]}^{'}$ $x_t = \left[ x^\prime_{m,t},x^\prime_{q,t}\right]^\prime$, where n_m×1 vector x_m,t collects variables observed at an annual frequency, such as the amount of fixed energy consumption and total foreign trade, and n_q×1 vector x_q,t collects unobserved yearly variables published at a quinquennial frequency, such as GDP. Definitions $X_{t} = {[x_{t}^{'}, \dots, x_{t - p + 1}^{'}]}^{'}$ $X_t = \left[ x^\prime_t, \ldots, x^\prime_{t-p+1}\right]^\prime$ and $Φ = {[Φ_{1}, \dots, Φ_{p}, Φ_{c}]}^{'}$ $\Phi = \left[\Phi_1, \ldots, \Phi_p, \Phi_c\right]^\prime$. Write the VAR in the concomitant form as 2 $X_{t} = F_{1} (Φ) X_{t - 1} + F_{c} (Φ) + v_{t}, v_{t} ~ i i d N (0, Ω (Σ))$ ${X_t} = {F_1}(\Phi ){X_{t - 1}} + {F_c}(\Phi ) + {v_t},{v_t}\sim iidN(0,\Omega (\Sigma ))$

where the first n rows of F₁(Φ)X_t−1, F_c(Φ), and v_t generate equation (1), and the remaining rows are defined as constant equations x_q,t−l = x_q,t−l, l = 1,…, p−1. The n×n upper-left submatrix of Ω is equal to Σ, and all other elements are 0. Equation (2) is the state-transfer equation for the MF-VAR.

The measurement equations are described next. There are several ways to deal with unobserved variables, for example, by entering 0 and modifying the measurement equations by setting the loadings on the state variables to 0; or by setting the measurement error variance to infinity; or by varying the dimensions of the observable vectors as a function of time t, the last of which is used in this paper to deal with unobserved variables. Let T denote the prediction origin and let T_b ≤ T be the last period corresponding to the last year of the quinquennium for which all quinquennial observations are available. Subscript b represents the equilibrium sample. The vector of yearly sequences x_m,t is observed for each month prior to period T_b. Denote the actual observations by y_m,t and write 3 $y_{m, t} = x_{m, t}, t = 1, \dots, T_{b}$ ${y_{m,t}} = {x_{m,t}},t = 1, \ldots ,{T_b}$

Assuming that the base year VAR has at least three lags, i.e., p ≥ 3, denote the three-month average of x_q,t as: 4 ${\tilde{y}}_{q, t} = \frac{1}{3} (x_{q, t} + x_{q, t - 1} + x_{q, t - 2}) = Λ_{q X} X_{t}$ $\tilde y_{q,t} = \frac{1}{3} \left(x_{q,t} + x_{q,t-1} + x_{q,t-2} \right) = \Lambda_{qX} X_t$

For variable ln GDP, measured in logarithms, the formula can be approximated as the arithmetic mean of GDP, which preserves the linear structure of the state space model. Let M_q,t be a selection matrix that equals the unit matrix if t corresponds to the last year of a quinquennium and is empty otherwise. 5 $y_{q, t} = M_{q, t} {\tilde{y}}_{q, t} = M_{q, t} Λ_{q X} X_{t}, t = 1, \dots, T_{b}$ ${y_{q,t}} = {M_{q,t}}{\tilde y_{q,t}} = {M_{q,t}}{\Lambda_{qX}}{X_t},t = 1, \ldots ,{T_b}$

For period t = T_b + 1,⋯,T, there are no additional observations available for the quinquennial time series. For these periods, y_q,t has dimension 0 and the selection matrix M_q,t in equation (5) is empty. However, additional yearly variables may be observed in this paper. Let y_m,t denote the subset of annual variables for which statistical agencies report observations for period t after period T_b, and let M_m,t be a deterministic sequence of selection matrices such that equation (3) can be expanded as 6 $y_{m, t} = M_{m, t} x_{m, t}, t = T_{b} + 1, \dots, T$ ${y_{m,t}} = {M_{m,t}}{x_{m,t}},t = {T_b} + 1, \cdots ,T$

Here the dimension of vector y_m,t is to be possible over time and is less than n_m. The measurement equations (3) to (6) can be simplified and written as: 7 $y_{t} = M_{t} Λ_{X} X_{t}, t = 1, \dots, T$ ${y_t} = {M_t}{\Lambda_X}{X_t},t = 1, \ldots ,T$

Here, M_t is a sequence of selection matrices that selects time t variables that are observed in period T and are part of the predictor’s information set. The state-space representation of the MF-VAR is given by equations (2) and (7).

3

Theoretical basis for synergistic research on carbon mechanisms and carbon tariffs

The problem of negative externalities of carbon emissions has prompted countries to explore effective emission reduction mechanisms. At the same time, carbon tariffs, as a kind of carbon emission reduction push mechanism, have profoundly influenced the practice of low-carbon economic transformation in various countries. However, the implementation of carbon tariffs has also triggered the concern of developing countries about trade protectionism. Studying the synergy between the carbon mechanism and carbon tariffs will help China reach its carbon emission reduction targets, maintain its advantages in foreign trade, and stabilize its total exports. This part will explore the synergistic emission reduction mechanism of carbon mechanism and carbon tariff and its necessity from the theoretical level.

3.1

Theoretical Basis for Synergistic Emission Reduction of Carbon Tax and Carbon Emission Trading

Carbon emissions have a strong negative externality, market players in the production or consumption of carbon dioxide emissions into the air, deteriorating air quality, but they do not have to bear any cost of pollution control, that is, the external cost, which makes the carbon emissions behavior of market players are not subject to any constraints. In the long run, it will not only damage the ecological environment, but also is not conducive to the sustainable development of the economy and society. Theoretically, there are various measures to correct the negative externality, such as Peguy tax, legal means, Coase’s theorem, integration and so on. In solving the pollution problem of carbon emissions, the most widely used in practice is the levy of carbon tax and carbon emissions trading.

3.1.1

Peguy’s theory of taxation

The rationale for a carbon tax is the Peguy tax. The tax on polluters is equivalent to the external cost, so that the external cost can be transferred to the internal private cost of polluters, and the polluters will spontaneously reduce the pollution behavior for less tax or no tax. It can be seen that the mechanism of levying carbon tax is to increase the cost of carbon emission subject through the form of tax, so as to make it reduce carbon emission voluntarily. Carbon tax as a carbon pricing mechanism, its tax rate level determines the carbon price and cost, that is, indirectly regulating carbon emissions in the form of tax rate setting carbon price. For the main body of carbon emission, the cost to be borne can be estimated in advance from the relatively determined level of tax rate, and reasonable production and operation decisions can be made to reduce emissions. However, if the tax rate is too low, the carbon emission subject as an “economist” may weigh the costs and benefits, resulting in the government’s pre-set emission reduction targets not being realized. It can be seen that there is great uncertainty in the emission reduction effect of the carbon tax, and the carbon emission reduction target cannot be quantified in advance.

3.1.2

The Coase Theorem

The theoretical basis for carbon emissions trading stems from the Coase Theorem. Under the premise of clear property rights and zero or small transaction costs, Pareto optimization can be achieved through market trading, which is the Coase Theorem. Further development of the Coase Theorem, put forward the theory of emissions trading, for carbon emission reduction problem provides a market solution mechanism. Carbon emission right is the emission right, which is also the core element emphasized by the Coase Theorem - property right, i.e., the government only needs to stipulate a total amount of carbon emission and issue the carbon emission right to the enterprises, then the enterprises can trade the carbon emission right in the market according to their actual carbon emission. The mechanism of carbon emissions trading is to treat carbon emission rights as a commodity, and enterprises with a shortage of carbon emissions can buy them from enterprises with a surplus, and the price of carbon is determined by the supply and demand mechanism in the market, which is in fact the cost of carbon emissions by enterprises. Since the total amount of carbon emissions is fixed, carbon emissions trading can achieve the government’s pre-set emission reduction targets, but the price will fluctuate greatly due to supply and demand, so enterprises cannot measure the cost in advance.

3.2

Two Carbon Emission Reduction Forcing Mechanisms of Carbon Tariffs

The moment the carbon tariffs are directed at large developing countries, the carbon emission reduction push mechanism for large developing countries has already begun to play its role. Taking whether the carbon tariff is implemented as the distinguishing point, the carbon emission reduction forcing mechanism can be divided into forcing active carbon emission reduction mechanism and forcing passive carbon emission reduction mechanism. Figure 1 is a diagram of the carbon reduction push mechanism formed by carbon tariffs. When the carbon tariffs have not really implemented, developed countries to force large developing countries to reduce carbon emissions can only be through international climate negotiations to pressure, in all comply with the “common but differentiated responsibilities” principle, the developed countries are not likely to reduce emissions targets imposed on the developing countries, but in the negotiation of a variety of diplomatic means to isolate the large developing countries, so that they actively commit to voluntary emission reduction mechanism. Instead, they will use various diplomatic means to isolate developing countries in the negotiations, so that they will take the initiative to commit themselves to voluntary emission reduction targets, that is to say, to force the active carbon emission reduction mechanism. If large developing countries adopt an uncooperative attitude in international climate negotiations, the deteriorating negotiation situation will make the implementation of carbon tariffs more likely. Once the developed countries to implement carbon tariffs, it means that through the international trade route, regardless of the willingness of large developing countries, will be pulled into the ranks of emission reduction, because do not reduce emissions means that exports of products with high implicit carbon content, the cost of carbon is internalized is also large, will inevitably lose the original foreign markets, and in order to make the export price is competitive, you need to carry out carbon emission reduction. This kind of carbon emission reduction forcing mechanism for large developing countries appear to be somewhat passive, can be called forcing passive carbon emission reduction mechanism. Whether it is a reverse active carbon emission reduction mechanism or a reverse passive carbon emission reduction mechanism, both imply that developing countries in the economic growth, but also need to carry out low-carbon development transition, the difference between the two mechanisms is only whether or not to give developing countries sufficient preparation to carry out the transition. In contrast, although the reverse active carbon emission reduction mechanism has been under pressure before the implementation of carbon tariffs, this pressure can also be transformed into motivation, so that large developing countries can independently decide on the time and process of transformation according to their own national conditions, which is a more desirable kind of reverse mechanism.

4

Study on the impact of carbon tariffs on the development of foreign trade in a low-carbon economy

This chapter constructs an MF-VAR model containing trade intensity, per capita energy consumption and per capita carbon emissions, and conducts an empirical study with actual data. Through the econometric steps of descriptive analysis of sample data, unit root test, cointegration analysis, causality test, impulse response analysis and variance decomposition, the impact between carbon tariffs and foreign trade in a low-carbon economy is systematically explored.

4.1

Modeling and Research Methodology

4.1.1

Model setup

The study of the relationship between foreign trade and carbon emissions in province A is taken as an example to explore the impact of carbon tariffs on foreign trade in a low-carbon economy. Observing the source of economic growth, it can be seen that energy consumption is an important factor affecting the growth of carbon emissions, so in order to enhance the accuracy of the empirical research results, this paper introduces energy consumption into the MF-VAR model. This paper argues that there exists a long-term equilibrium relationship between energy consumption, foreign trade and carbon emissions. This paper takes trade density as the indicator of foreign trade, and takes per capita energy consumption and per capita carbon emissions as the indicators of energy consumption and carbon emissions, and constructs the following model: 8 $C A = f (E N, T)$ $CA = f\left(EN, T\right )$

Assuming that the relationship between carbon emissions and energy consumption and foreign trade presents a linear relationship, taking logarithmic values for variables CA_n and EN_n eliminates the exponential growth trend of carbon emissions li energy consumption, the specific form of which can be written as: 9 $L n C A A_{n} = α + β_{1} L n E N_{n} + β_{2} T_{n} + u_{n}$ $LnCAA_n = \alpha + \beta_1 LnEN_n + \beta_2 T_n + u_n$

In the above equation (9), CA_n represents the per capita carbon emission of province A in period n; EN_n represents the per capita energy consumption of province A in period n; T_n represents the degree of openness to the outside world; β₁ and β₂ are the unknown parameters, and u_n is the random error term.

It is well known that carbon emissions increase with rising energy consumption. Therefore, the prediction β₁ is positive. According to the assertion of the EKC curve, if the estimated sign of β₂ is positive, it means that foreign trade makes the carbon emission of province A increase; if it is negative, it means that foreign trade favors province A’s provincial energy emission reduction. Since foreign trade in province A is high energy consumption, high emission and high pollution, we theoretically expect β₂ to be positive.

4.1.2

Research methodology

1)

Descriptive analysis of the sample. Make a growth trend graph of LnCA_n, LnEN_n and T_n to analyze the linear relationship and smoothness between the three.

2)

Unit root test. Use the extended Dick Fuller test to test the smoothness of the series LnCA_n, LnEN_n and T_n.

3)

Cointegration analysis. Cointegration mainly reflects the existence of a long-term stable relationship between non-stationary single variables. Cointegration test commonly used methods are Engel-Granger two-root test and Johansen test, Engel-Granger test is usually used for the test of cointegration between two variables, for the test of cointegration between multivariate variables, you can use the Johansen test based on the vector autoregressive model. In this paper, the Johansen method is used to test whether there is a long-term equilibrium relationship between LnCA_n, LnEN_n, T_n.

4)

Causality test. Through the cointegration test, it shows that there is a cointegration relationship between carbon emissions, foreign trade and energy consumption in province A. However, what kind of causality exists in this long-term equilibrium relationship needs to be tested by Granger causality test for LnCA_n, LnEN_n and T_n.

5)

Impulse response analysis. The impulse response model is used to analyze the dynamic feedback relationship between LnCA_n, LnEN_n and T_n.

6)

Variance decomposition. Variance decomposition is usually used to measure the contribution of the impact factors to the explanatory variables by the impact of the factors, through the VAR model to dynamically represent the contribution of the low-carbon economy to foreign trade.

4.2

Measurement process and analysis of results

4.2.1

Descriptive analysis of sample data

In the past three decades, foreign trade in Province A has made rapid progress. The total value of import and export was only 3.53 billion dollars in 1985 to 188.96 billion dollars in 2011, ranking seventh in the country. At the same time, along with the rapid growth of foreign trade, energy consumption and carbon emissions have also increased. per capita energy consumption in 1985 was only 0.69 tons of standard coal, and per capita energy consumption in 2011 amounted to 2.62 tons of standard coal, an increase of nearly 3.8 times. per capita carbon emissions in 1981 amounted to 0.607 tons of standard coal, and per capita carbon emissions in 2011 amounted to 2.614 tons of standard coal, an increase of nearly 4 times, creating a high level of carbon emissions. The total per capita carbon emissions in Province A amounted to 2.614 tons of standard coal in 2011, an increase of nearly four times, creating a situation of high pollution, high energy consumption and high emissions.

Fig. 2 shows the growth trend plots for LnEN_n, LnCA_n and T_n. Figure 3 is the second order difference plot for LnEN_n, LnCA_n and T_n. From Figures 2 and 3, it can be seen that the original series LnCA_n, LnEN_n, and T_n are non-stationary with fluctuation range levels between -0.9 and 1.5 during 1981-2011, with large fluctuations. While the second-order difference series of LnCA_n, LnEN_n and T_n are smooth, with fluctuation range level between -0.5-0.4 and less fluctuation. Through the second-order difference series, it can be judged that there is a long-term equilibrium relationship between LnCA_n, LnEN_n and T_n.

4.2.2

Unit root test

MF-VAR model econometric analysis requires smoothness of data variables. The method of econometric analysis is valid only if the variables in the model satisfy the smoothness requirement. If the model contains non-stationary time series, the conclusions of statistical estimation and testing may be wrong and the problem of spurious regression occurs. For non-smooth variables, there may be cointegration between the variables only if the single integer order of the explanatory variables is not lower than the single integer order of the explanatory variables. Therefore, when constructing the MF-VAR model, the smoothness of the time series should be tested first and the single integer order of the variables should be further determined before the MF-VAR model can be built and the cointegration analysis can be performed. If the variables in the MF-VAR model have unit roots, it means that the time series data are non-stationary series. However, if the first-order difference series of the non-stationary variable is smooth, i.e., first-order single-integration I(l), the VAR model can also be constructed using the difference form of the variable.

In this study, the econometric software Eviews 6.0 is applied to determine the smoothness of the time series LNCT1, LNCT2, and LNEX by applying the ADF unit root test for level and first-order difference unit root tests, respectively.

Table 1 shows the results of the unit root test under the ADF method. The presence of unit root in the horizontal time series of LNCT1, LNCT2 and LNEX respectively proves that all the three time series are non-stationary. However, the hypothesis of the existence of unit root is rejected at 5% level of significance for the first order difference series D(LNCT1), D(LNCT2) and at 1% level of significance for D(LNEX). Therefore, the first order difference series of LNCT1, LNCT2 and LNEX are all smooth, so these variables are I(1) single-integrated series, which satisfy the prerequisites of the cointegration test, and can be utilized to conduct Johanson cointegration test and analyze the long-run relationship between the variables using cointegration equations.

Table 1.

Results of unit root test of ADF method

Variable	ADF value	Significant level	Threshold	Associated probability	Conclusion
LNEX	-0.906425	1% level	-3.80854	0.7646	No
		5% level	-3.02068
		10% level	-2.6504
D(LNEX)	-4.411628	1% level	-3.8315	0.0031	Yes***
		5% level	-3.02996
		10% level	-2.65518
LNCT1	-1.689862	1% level	-3.80854	0.4210	No
		5% level	-3.02068
		10% level	-2.6504
D(LNCT1)	-3.823724	1% level	-3.8315	0.0103	Yes**
		5% level	-3.02996
		10% level	-2.65518
LNCT2	-1.684928	1% level	-3.80854	0.4233	No
		5% level	-3.02068
		10% level	2.65042
D(LNCT2)	-3.786470	1% level	-3.8315	0.0111	Yes**
		5% level	-3.02996
		10% level	-2.65518

Note: D represents the first difference of the variable;

means rejection of null hypothesis at 1% significance level;

means rejection of null hypothesis at 5% significance level;

means that the null hypothesis is rejected at a significance level of 10%

4.2.3

Cointegration tests

Cointegration theory analyzes the long-run equilibrium relationship between non-stationary economic variables, and the cointegration test can be used to determine whether there is a cointegration relationship between the variables and each other. If the cointegration test determines that there is a cointegration relationship between the variables, i.e., a long-run equilibrium relationship, the growth rates of the variables show that they have a common growth trend.

The variables LNEX and LNCT1 are subjected to the johanson cointegration test, and Table 2 shows the results of the cointegration test between the volume of foreign trade in industrial goods and carbon tariffs. The results of the cointegration test in Table 2 show that there is 1 cointegration relationship between variables LNEX and LNCT1 at 5% significance level. It indicates that there is a long-run equilibrium relationship between variables LNEX and LNCT1 during the sample period. The expression of the cointegration relationship is: LnEX + 3.68LnCT1 = 0. From the cointegration equation, it can be seen that the relationship between the volume of China’s foreign trade in industrial goods and carbon tariffs is negatively correlated, and the average elasticity of the export volume with respect to the carbon tariffs is 3.68. That is to say, for every 1% increase in the logarithm of the carbon tariffs, the logarithm of China’s foreign trade in industrial goods will be reduced by 3.68% on average. Since China is the world’s top exporter and the base of China’s foreign trade in industrial goods is large, it can be predicted that the absolute value of the reduction in China’s foreign trade in industrial goods is quite large. By extension, the imposition of carbon tariffs by developed countries will deal a heavy blow to the foreign trade of developing countries, including China.

Table 2.

Test results of industrial exports and carbon tariff co-integration

Hypothesized No. of CE(s)	Eigenvalue	Trace Statistic		0.05 Critical Value	Prob.**
None *	0.529117	15.66486		15.49470	0.0470
At most 1	0.110523	2.108218		3.841465	0.1464
Normalized cointegrating coefficients (standard error in parentheses)
LNEX			LNCT1
1.000000			3.686451(0.41359)
Adjustment coefficients (standard error in parentheses)
D(LNEX)			0.190010(0.07245)
D(LNCT1)			0.221410(0.06363)

4.2.4

Granger causality test

Since the cointegration relationship can only illustrate the long-term relationship and development trend between the variables, it is necessary to further analyze the causal relationship between the variables through Granger causality test, which refers to whether a variable is helpful to the prediction of another variable, and the Granger causality test is very sensitive to the smoothness of the variables. From the above cointegration analysis, it can be seen that there is a long-term equilibrium relationship between China’s foreign trade in industrial goods and carbon tariff. But in the short run, is there a causal relationship between the two, and what is the direction of causality? Hypothesis: In the short run, there is no causal relationship between the two.

Granger causality test requires that the time series must be smooth, although LNCT1, LNCT2, LNEX are non-smooth series, but according to the previous ADF unit root smoothness test results, it can be determined that their first-order difference series are all smooth series, so LNCT1, LNCT2, LNEX meets the prerequisites of the Granger causality test and can be be analyzed by Granger causality test. Table 3 shows the results of Granger causality test of industrial export volume and carbon tariff. According to the Granger causality test results in Table 3, it can be seen that the original hypothesis is accepted because the p-value of foreign trade in manufactured goods is not a Granger cause of carbon tariffs is 0.2202, which is greater than 0.05. The P-value of carbon tariffs not being a Granger cause of foreign trade in manufactured goods is 0.0486, which is less than 0.05, so the original hypothesis is rejected, so carbon tariffs are a Granger cause of China’s foreign trade in manufactured goods, and the levying of carbon tariffs will reduce the amount of exports of manufactured goods. This indicates that China’s current foreign trade in industrial products is characterized by crude growth, which is an energy-consuming pattern of high energy consumption, high pollution and high emission.

Table 3.

Granger causality test between industrial exports and carbon tariffs

Null Hypothesis:	Obs	F-Statistic	Prob.
LNCT1 does not Granger Cause LNEX	20	3.77845	0.0486
LNEX does not Granger Cause LNCT1		1.61949	0.2202

4.2.5

Impulse response

The impulse response function mainly describes how the current and future values of the endogenous variables are affected when the error term is increased by a one-unit (standard deviation) shock. Under the influence of carbon mechanisms and carbon tariffs, China needs to carry out a low-carbon economic transformation, and the low-carbon economic transformation will definitely affect China’s foreign trade in the long or short term. This paper applies the impulse response function method to analyze the impact of low-carbon economy on China’s foreign trade. Impulse response analysis is carried out on the basis of the constructed MF-VAR model, and the impulse response graph of low carbon economy development level C on foreign trade T is obtained.

Figure 4 shows the response of foreign trade to one standard message. Observing Fig. 4, foreign trade has a downward trend from the 1st period and reaches the lowest value in the 3rd period when it is hit by the one standard information of low carbon economy. It indicates that the low-carbon economy has a negative feedback effect on foreign trade in this period. After that, it gradually rises from the 4th period, and the rate of impact is increasing and reaches the maximum rate of impact in the 8th period. From the 8th period onwards, although the development of low-carbon economy still has an impact on foreign trade, the rate of impact gradually slows down. The reason for the negative impact in the 1st period and reaching the negative maximum in the 3rd period is that: in the early stage of the development of low-carbon economy, China’s low-carbon products may not be perfect and therefore not recognized by foreign countries, thus it will have a negative impact on foreign trade and reduce the total amount of foreign trade; after that, in the beginning of the 4th period, the impact impact influence of the low-carbon economy on foreign trade begins to rise gradually, which indicates that with the passage of time This shows that with the passage of time, because of the continuous development of China’s low-carbon economy, the foreign trade products will be upgraded and perfected, and foreign countries will recognize and trust China’s low-carbon products, so the total amount of foreign trade will increase.

4.2.6

Variance decomposition

Variance decomposition is usually used to measure the degree of contribution to the explanatory variables by the impact of the influencing factors by the impact of the MF-VAR model to dynamically represent the contribution of the explanatory variables to the explanatory variables, and Table 4 shows the results of the variance decomposition of the low-carbon economy on foreign trade. According to Table 4, the contribution of foreign trade to itself is 100% in the 1st period, and after that the contribution is decreasing steadily and reaches a steady state around the 8th period, and its contribution is about 74.5% or so. The level of low-carbon economic development has no effect on foreign trade in the 1st period, and the contribution rate starts to rise gradually from the 3rd period, and the rate of increase is relatively stable, and reaches a steady state around the 8th period, and the contribution rate is about 26.2%. This shows that, forced by the carbon mechanism and carbon tariffs, the transformation and development of low-carbon economy has a significant impact on foreign trade. With the passage of time, the impact of the level of low-carbon economic transformation and development on foreign trade is gradually increasing. In the short term, the contribution rate of low-carbon economic transformation to foreign trade is about 16.1%, and in the long term, it will finally stabilize and reach a contribution rate of 26.2%. Therefore, the impact of the development level of low-carbon economic transformation on foreign trade is enormous. To increase China’s total foreign trade in the future, vigorously developing the low-carbon economy is an aspect worth considering.

Table 4.

Variance decomposition of foreign trade

Period	T(%)	C(%)
1	100	0.00000
2	93.13578	6.86422
3	85.19534	14.80466
4	83.85902	16.14098
5	79.95932	20.04068
6	76.75789	23.24211
7	74.77952	25.22048
8	74.56891	26.23109
9	73.37278	26.62722
10	73.18066	26.81934

5

Conclusion

This paper constructs mixed-frequency vector autoregressive model by combining trade intensity, per capita energy consumption, and per capita carbon emissions to study the impact of carbon tariffs on foreign trade as well as the dynamic relationship between low-carbon economic transformation and foreign trade. In the descriptive analysis of the sample data, the fluctuation range levels of the second-order difference series of the three variables of energy consumption, carbon emissions and time are between -0.5 and 0.4, and there exists a long-run equilibrium relationship among the three. The horizontal time series of LNCT1, LNCT2 and LNEX reject the hypothesis of the existence of a unit root under the significance of 5% and 1%, respectively, which is sufficient to fulfill the prerequisite conditions of cointegration test. The cointegration test reveals that the relationship between foreign trade volume and carbon tariffs is negatively correlated. The average elasticity of export volume to carbon tariff is 3.68. Combined with the Granger causality test, since the p-value of carbon tariff is not a Granger cause for foreign trade of industrial manufactured goods is less than 0.05, it can be understood that China’s current industrial goods with high energy consumption, pollution, and carbon emissions will be affected by the decline in trade volume brought about by carbon tariffs, and so on, and it is necessary to take a series of measures to intervene.

Based on the negative impacts of carbon mechanism and carbon tariffs on China’s foreign trade, it is chosen to transform from a rough economy to a low-carbon economy. The impact of low-carbon economic transformation on China’s foreign trade is analyzed through the impulse response, and it is found that the feedback of low-carbon economic transformation on China’s foreign trade is negative in the phase of 1-3, and the positive feedback can only be obtained gradually from the phase of 4. Combined with the variance decomposition to understand the contribution rate of low-carbon economic transformation to foreign trade. In the short term, the contribution rate is only about 16.1%, but with the success of the transition, the contribution rate of about 26.2% can be obtained stably in the end. This shows that in the process of low-carbon economy transition, policymakers should fully consider the insufficient contribution rate in the early stage, and construct a reasonable synergistic policy framework of carbon mechanism and carbon tariff, so as to help industries or enterprises willing to transition to a low-carbon economy to pass through this stage, and then obtain stable foreign trade gains afterwards, in order to realize a win-win situation for both the economy and the environment. In the meantime, China’s foreign trade competitiveness will be enhanced while coping with climate change.

Sprache:: Englisch

Zeitrahmen der Veröffentlichung:: 1 Hefte pro Jahr
Fachgebiete der Zeitschrift:: Biologie, Biologie, andere, Mathematik, Angewandte Mathematik, Mathematik, Allgemeines, Physik, Physik, andere

Zeitschrift RSS Feed

Synergistic policy construction of carbon mechanisms and carbon tariffs in the transition to a low-carbon economy

Yinglan Jin

Online veröffentlicht: 25. Sept. 2025

Eingereicht: 15. Jan. 2025

Akzeptiert: 30. Apr. 2025

DOI: https://doi.org/10.2478/amns-2025-1026

SchlüsselwörterLow carbon economy, Carbon mechanism, Carbon tariff, MF-VAR model, Econometric analysis, Foreign trade

© 2025 Yinglan Jin, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Schlüsselwörter
Low carbon economy, Carbon mechanism, Carbon tariff, MF-VAR model, Econometric analysis, Foreign trade