Calculation of GTFP and its influencing factors in Guangdong-Hong Kong-Macao Greater Bay Area

: Under the background of dual carbon target, the manufacturing industry in the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) is facing the pressure of green transformation. This paper takes this as the research basis, collects the production panel data of the manufacturing industry in the GBA in the past decade, calculates the green development level of the GBA and finds out the important influencing factors, so as to provide theoretical suggestions for the green production and green development of the manufacturing industry in the GBA.


Introduction
The Guangdong-Hong Kong-Macao Greater Bay Area (GBA) is a new force in the world's bay area economies and has increasingly become a bright spot in the world's economic landscape. At present, although the manufacturing industry in the GBA has made some achievements in the development process, it is still faced with problems such as the traditional extensive development mode has not changed fundamentally, the lack of innovation motivation of small and medium-sized enterprises, and the lack of green production awareness of enterprises, which seriously restrict the development of the GBA. [1]

Literature review
The evaluation of low-carbon and green development is the basis of low-carbon and green transformation of manufacturing industry. Scholars have been constantly improving the evaluation system and calculation model. In the early stage of the study, scholars mainly adopted green GDP and total factor productivity (TFP). Green GDP was obtained by deducting resource consumption and environmental loss from GDP, but it was difficult to measure environmental loss in actual conditions. The main factors considered in the calculation of TFP are labor and capital input, and the data is easy to obtain. However, the constraints of environmental ecological factors and resource shortage on economic development are not taken into account, which is inconsistent with the actual situation. Later, scholars gradually adopted the index of green Total factor productivity (GTFP). Li Jun and Xu Jintao (2009) [2] added variables reflecting environmental change into the calculation of TFP and took it as an indicator of economic growth, and for the first time applied the concept of green total factor productivity (GTFP) in domestic literature. There are parametric methods and non-parametric methods to measure GTFP. When using parameter method, GTFP should be decomposed into technical progress, technical efficiency, scale effect, etc. Wu Yiqing and Chen Junxian (2018) [3] used the parameter method to estimate GTFP in Hebei. Nonparametric method mainly uses data envelopment analysis (DEA) method to calculate. In addition, scholar Li Lin et al. (2018) introduced the unexpected output into the directional distance function to construct the Malmquist-Luenberger green total factor productivity index. In addition, some scholars introduced the relaxation amount of input and output to estimate the direction distance function and build the (SBM) model. [4] The above methods all have certain defects, for example, ML index is highly sensitive to the sample period, and SBM model has time lag for calculating R&D inputoutput, etc. According to the literature review, there are few relevant papers on the measurement of green TFP in the Guangdong-Hong Kong-Macao Greater Bay Area. Li Ping and Ju Chuanxiao (2022) [5] studied the comparison of green innovation development between the Yangtze River Delta region and the Guangdong-Hong Kong-Macao Greater Bay Area, but they used the ML model, which could not overcome the sensitivity of the selection period. Zhang et al. (2019) [6] calculated the green competition index of manufacturing industry in the Guangdong-Hong Kong-Macao Greater Bay Area, which mainly measured the regional differences. Therefore, based on this, this paper selects the production data of nine manufacturing cities in Hong Kong, Macao and Guangdong Province from 2011 to 2020 and uses the panel data model to calculate the green competition index. In this paper, the SBM-GML model is used to improve the calculation model, further optimize the model, and make the calculation results more realistic and comparable. In addition, the paper innovatively introduces unexpected output to make the measured results more authentic and credible. Finally, this paper innovatively takes industry as the classification standard to deeply analyze the impact of green total factor productivity caused by industry differences, which is of more practical significance.

UE
(1) Therefore, the pollution intensity coefficients of different manufacturing industries in the GBA from 2011 to 2020 are obtained, as shown in the following table. Table 1 Pollution intensity of manufacturing industries in the GBA The following is a breakdown of industries based on pollution intensity.

Model construction
This paper uses SBM-GML model to measure green TFP. Green total factor productivity (GTFP) is based on total factor productivity (TFP), which can reflect the environmental factors such as energy consumption and environmental pollution level. The SBM model is shown below.
And then construct the SBM-GML model (from period t to period t+1) as follows.  4.2 Measure the exponential decomposition Chung (1997) [9] concluded in his study that GML index is composed of EC index (technical efficiencychange) and TC index (technological change, production technology change), which has been widely recognized in the academic community. This conclusion can be used to measure GTFP.

Analysis of GTFP (Green Total factor Productivity) measurement results
In this paper, Matlab software is used for calculation, and the calculation results are shown below. According to the analysis of the above indicators, from 2011 to 2020, the average annual growth rate of green TFP in the GBA is -0.4%, the growth rate of green technical efficiency is 1.2%, and the growth rate of green technical progress is -1.6%. The data show that the green TML level in the Bay Area is facing a certain development bottleneck. The green technical efficiency has a relatively fast growth rate, while the development trend of green technical progress efficiency is relatively low. This shows that the management level and industrial structure of GBA have been further improved and become more reasonable, and the growth rate of green technology efficiency has been steadily increasing. However, due to the influence of environmental policy restrictions and industrial transformation and upgrading, the expected output has decreased under the same input level, but the overall production level is approaching to the maximum benefit, and the allocation of production and output is good.

Construction of the index system of influencing factors
Other indicators adopt the indicators recognized by scholars, for environmental regulation. The GBA mainly implements economic incentive environmental regulation. In order to avoid the overlap with the previous classification standard of industrial pollution degree, this paper chooses the research method of Huang Qinghua (2018) [12] and uses the operating cost of pollution control facilities per unit of pollution emission as the quantitative index. The specific calculation formula is as follows.

ER
C P 11 Table 5 Selection of influencing factors and indicators of green TFP

Model setting
This paper collects the production data of the GBA manufacturing industry from 2011 to 2020, selects 26 industries, and constructs static and dynamic panel data models.

Static panel model
The formula is as follows. lnGTFP a a lnCL a lnSC a lnTE a lnFDI a lnER ε 12 lnGEC a a lnCL a lnSC a lnTE a lnFDI a lnER ε 13 lnGTC a a lnCL a lnSC a lnTE a lnFDI a lnER ε 14

(4) Data test
First, the raw data were processed and the variables were log processed. The results of descriptive statistics on the data are shown in the following table.  For both static and dynamic multicollinearity tests, the VIF and its mean are small and much less than 10, and the mean VIF is 1, so there is no collinearity problem. The data were tested for collinearity, and the results are The model test results show that the mixed regression model is better than the fixed effect model, the LM test shows that the mixed regression model is better than the random effect regression model, and the Hausman test shows that the random effect model is better than the fixed effect model. In summary, the mixed regression model is selected for analysis. The static panel regression analysis is carried out below. Models 1-3 represent the regression analysis with GTFP, GEC and GTC as explanatory variables, and the results are shown in the following table. Table 9 Results of static panel analysis

. Multicollinearity test
The regression results show that in model1, the degree of capital deepening (CL) has a negative impact on the GTFP of the manufacturing industry in the GBA, and H1 is established. In Model 1-3, industry average size (SC) has a positive impact on manufacturing GTFP in GBA, and also has a positive impact on GEC and GTC, H2 is established. In Model 1-3, the degree of technological innovation has a negative impact on GTFP, GEC and GTC of the manufacturing industry in the GBA, which is inconsistent with H3. It is assumed that due to the mismatch between input and output, technological innovation needs to invest a lot of capital in the early stage, but not all the investment can be transformed into green technology achievements, and the transformation of scientific research achievements into technology needs to go through multiple audits, and the time span is long, so the overall contribution to green production is not obvious for the time being. In Model 1-3, the level of foreign direct investment (FDI) has a negative impact on the manufacturing GTFP, GEC and GTC in the GBA, which is inconsistent with H4. It is speculated that the opening to the outside world will also lead foreign investors to bring heavy polluting enterprises to the Bay Area. In model1 and model3, the environmental regulation level (ER) has a positive impact on GTFP and GTC of the manufacturing industry in the GBA, and a negative impact on GEC of the manufacturing industry in the GBA in model2. It is consistent with H5. It is speculated that environmental regulation makes the manufacturing industry in the Bay Area pay more attention to technological progress and ignore green technology efficiency, which leads to a negative impact on GEC.
5 Research conclusions and suggestions 1. Clean energy consumption structure The GBA should change the energy structure dominated by coal, expand the proportion of clean energy, respond to the national dual carbon target, release the constraints of energy on the manufacturing industry in the Bay Area, and fundamentally improve the green development level of the manufacturing industry in the Bay Area.
(2) Increase investment in green production technology research and development It can be seen from the above research that technological progress is an important factor affecting the green production level of the manufacturing industry in the Bay Area. Therefore, the government should introduce relevant policies and encourage enterprises to increase investment in green production technology through carbon sink market, carbon finance and other means to enhance the competitiveness of green production.
(3) Create a new pattern of opening up and strictly supervise foreign investors The study finds that building an open pattern and introducing foreign investment have a positive impact on the improvement of green technology efficiency, especially for heavily polluting industries, but have a negative impact on the technological progress of heavily polluting industries. That is, the manufacturing industry in the Bay Area may rely too much on foreign technology while ignoring the improvement of its own green production technology. In addition, it may lead foreign investors to transfer heavy polluting enterprises to the Bay Area, which is not conducive to the development of enterprises and the economic development of the Bay Area. In addition, the large scale of foreign investment has a negative impact on the green TFP of medium and light polluting enterprises. Therefore, the Bay Area should strictly supervise and control the pollution emissions of manufacturing enterprises when building a pattern of opening to the outside world. (4) Subdividing industries and taking targeted measures Through the above research, it can be found that the heterogeneity of different industries in the manufacturing industry is strong, so the government cannot solve the problem by "one size fits all". It needs to classify different industries and implement targeted environmental regulation policies. For heavily polluting enterprises, their own low-carbon transformation has a long input cost and transformation time, so they need to formulate relatively loose policies. Moreover, low green technology efficiency is an important reason for their low green production level, so they need to solve the problems at the management level. For enterprises with moderate pollution and light pollution, the slow progress of green technology is the reason that hinders the development of their green production level. Therefore, it is necessary to encourage these enterprises to increase capital investment and accelerate the transformation from scientific research achievements to production technology.