Measuring the Green Total Factor Productivity in Chinese Aquaculture: A Zofio Index Decomposition

Abstract

Aquatic products are important sources of protein and food consumption, which are pivotal to solving the problem of food shortage. As the world’s largest producer of aquatic products, China’s aquaculture has developed rapidly. However, a large quantity of pollutants has also been generated in the fish farming process, which is detrimental to the sustainable development of China’s aquaculture. Therefore, under double constraints with regard to resources and the environment, fishery development must achieve cleaner production. Measuring green total factor productivity in aquaculture is fundamental to improving aquaculture production efficiency and reducing pollution emissions. This paper investigated the green total factor productivity in China using the SBM-ML method and analyzed the dynamic evolution of fish farming, measuring its change characteristics with regard to time and spatial differences. The results indicated that the total factor productivity indexes of mariculture and freshwater aquaculture in China are 1.050 and 1.060, respectively. Regionally, mariculture in the East China Sea region has the highest green total factor productivity of 1.072, followed by the South China Sea region with 1.056, and the green total factor productivity in the Yellow Sea region is the lowest—only 1.030. The results of the freshwater aquaculture calculations showed the opposite geographical distribution characteristics in China, with the highest in the western region (1.098), followed by the central region (1.046), with the lowest being in the eastern region (1.038). Evidently, both mariculture and freshwater aquaculture have noticeable spatiotemporal heterogeneity. Relevant policy recommendations are provided to improve the green production efficiency of fishery.

1. Introduction

According to the latest global aquaculture statistics compiled by FAO (2019), aquatic products are high-quality protein sources. Fish consumption accounts for 17% of the overall global population’s animal protein intake, in comparison with 7% of ‘all’ protein consumption [1,2,3,4]. Moreover, China is the world’s major producer of aquatic products, accounting for 35% of global fish production in 2018. Aquaculture is mainly divided into inland freshwater aquaculture and coastal mariculture; however, the production of freshwater aquaculture is approximately six times greater than coastal mariculture. At the beginning of the 21st century, aquaculture production maintained rapid growth. In 2020, world aquaculture production reached a record high of 214 million tons, a slight increase of close to 3% from 213 million tons in 2018. Notably, from 2001 to 2018, global aquaculture production increased on average by 5.3% annually, while the growth rates in 2017 and 2018 were only 4% and 3.2%, respectively. Compared with 2019, world aquaculture production of animal species grew by 2.7% in 2020, the lowest annual growth rate in more than 40 years [5]. The slowdown in aquaculture production growth in China—the largest producer—is the main reason for the low growth of global aquaculture production in recent years. Therefore, it is imperative to study the production efficiency of aquaculture in China for the global aquatic product market [6].

“Blue Transformation” is a vision for sustainably transforming aquatic food systems. This is a recognized solution for ensuring food and nutrition security and environmental and social well-being by preserving aquatic ecosystem health; reducing pollution; protecting biodiversity; and promoting social equality [5]. For China, this action commenced with the strategy of sustainable development. This strategy was incorporated into China’s long-term economic and social development plan as early as 1994. China announced the goal of reducing carbon emissions, independently, within the next 20 years. At the 75th session of the United Nations General Assembly, China solemnly pledged to stabilize its carbon dioxide emissions by 2030 and achieve carbon neutrality by 2060. Under the constraint of the ‘double carbon’ goal, it is imperative to improve the green production capacity of aquaculture.

The growth rate of aquaculture production in China is strengthening; however, the resulting aquaculture pollution cannot be overlooked. Aquaculture is a biotransformation process that converts feed into animal products. Therefore, it is difficult to accumulate such pollutants as feed waste, solid waste, and soluble waste generated through conversion [7]. According to the latest national census of pollution sources, the total pollution emissions of aquaculture accounted for only 6.35% of the pollution from non-point agricultural sources, in which chemical oxygen demand, ammonia nitrogen, total nitrogen, and total phosphorus were 6.66 million tons, 2.23 million tons, 9.91 million tons, and 1.61 million tons, respectively. Increasing the production capacity of aquaculture to manage the challenges of climate and disease, on the premise of reducing pollutant emissions, is the prominent future development direction of aquaculture [8].

The resolute commitment to improving the green production efficiency of fish farming is a necessary means of mitigating environmental and resource constraints whilst also maintaining the sustained and sound development of the global economy and society. As a result of the multiple uncertainties caused by epidemics, wars, and natural disasters, the World Food Programme (WFP) has warned that humankind may face the greatest food crisis since the Second World War, with 1.7 billion people exposed to poverty and hunger as a consequence of deterioration in the food, energy, and financial system. The efficient development of fishery may prove to be one of the significant ways to solve the food shortage. As the largest developing country, China has a strong production capacity and consumption capacity for aquatic products. Studying China’s sustainable development path of aquaculture can provide valuable experience and effective methods for alleviating food shortages and for the sustainable development of fishery in developing countries.

This study has important theoretical significance, derived from a substantial number of studies that have focused on the production of aquatic products, which, however, did so without combining freshwater aquaculture with mariculture in China. Based on the panel data of mariculture and freshwater aquaculture in China from 2003 to 2020, the SBM-ML model and the Zofio index were used to calculate the green total factor productivity of aquaculture in China. Finally, this study presents policy suggestions for improving cleaner production efficiency.

2. Literature Review

Tinbergen (1942) initially proposed using the total factor productivity index to reflect the comprehensive utilization efficiency of input factors. Solow (1957) studied the production function and discovered that, in addition to the contribution of labor and capital, there remains the role of technological progress, which is entitled total factor productivity [9]. Squire (1992) originally applied TFP to fisheries, which rendered fishery research methods and systems more scientific and provided strong support to many scholars studying fishery development [10]. Stochastic frontier analysis (SFA) and data envelope analysis (DEA) are the two most commonly used methods for measuring total factor productivity in the academic circle [11].

Aigner et al. (1977) suggested the stochastic frontier approach (SFA) as the causal defect of Solow’s value method in measuring total factor productivity [12]. Battese and Coelli (1995) optimized the SFA in their study and measured the growth structure by econometric methods [13]. Kumbhakar and Lovell (2000) further apportioned total factor productivity into technical progress, technical efficiency, allocation efficiency, and scale efficiency [14]. However, the SFA has certain defects. It cannot reasonably control the relationship between input variables and output variables or solve endogenous problems if the wrong set of production functions would lead to a deviation of estimation results [15].

Sten (1953) originally proposed the Malmquist index method. On the basis of Sten’s research, Charnes et al. (1978) proposed the DEA–Malmquist index method to create a widely used method in the field of productivity measurement [16]. Chuang et al. (1997) constructed the ML productivity index based on the directional distance function to solve the shortcomings of the traditional TFP calculation method, which cannot consider the environmental pollution problem [13]. Thereafter, scholars have calculated green total factor productivity by considering undesirable production. Wang (2018) used the ML index method based on China’s provincial panel data to measure the total factor productivity of China’s transportation industry and discovered that China’s transportation industry demonstrated a downward trend due to energy and carbon emission constraints [17]. Based on the nonparametric method, the empirical results illustrated that the change in green total factor productivity is predominantly derived from the varying production efficiency in most provinces, rather than technological progress [18]. Since the Eleventh Five-Year Plan, the national average green total factor productivity growth has accelerated after readjusting the development direction in reducing pollution and conserving energy. Li (2021) applied the SBM-ML model to measure the green total factor productivity of urban agglomeration in the Pearl River Delta and discovered that the volatility of green total factor productivity in the Pearl River Delta region increased from 2005 to 2018 [19].

In the agricultural sector, Thijssen (1999) used pollution as an input factor to evaluate the technical and environmental efficiencies of dairy farms in the Netherlands. Zhong (2021) constructed a three-level MML index considering undesirable output based on the DDF-DEA model. It was discovered that, in terms of scale, the green total factor productivity of medium-scale laying hens in China was the highest, followed by the small-scale and large-scale. From a regional perspective, it was the highest in the western region, followed by the central and eastern regions [20]. However, the DDF model cannot calculate the arbitrary problems of insufficient improvement and poor output disposal [21]. Applying stochastic frontier analysis and the Malmquist index method to estimate the provincial GTFP in China from 2000 to 2016, one study determined that the overall trend was upward, with the highest GTFP in the eastern region and the lowest in the western region [22]. The Super-SBM model was used to calculate the agricultural green total factor productivity, based on carbon emissions of the main provinces and cities in China. Concurrently, the kernel density estimation method and the panel data model have been used to analyze dynamic evolution and the influencing factors of agricultural green total factor productivity in China [23]. The study discovered that China’s agricultural green total factor productivity indicated a fluctuating growth trend, and the disparities between provinces were growing. Agricultural factor endowments and regional characteristics affected China’s agricultural green total factor productivity, and these effects have regional differences.

There have been various studies on the efficiency of aquaculture. A Färe–Primont index approach was used to analyze the efficiency of aquaculture with regard to the ‘Gher’ aquaculture system in Bangladesh. It was discovered that the total factor productivity increased by 0.86% per year, mainly driven by technological changes of 0.54% per year and assorted efficiency changes of 0.06% per year [24]. Data envelopment analysis, combined with the double-bootstrapping technique, was used to analyze the cost and technical efficiencies of intensive shrimp farming in Ningshun, Vietnam. The study revealed that vast improvement was required with regard to the cost and technical efficiencies of intensive whiteleg shrimp farming [25]. Asche et al. (2013) analyzed the change of total factor productivity in the salmon farming department of Norway from 1996 to 2008 by applying the Malmquist index method, stochastic frontier analysis (SFA), and data envelopment analysis (DEA) [26]. The study demonstrated that the total factor productivity changed by 1–2% annually, in which the contribution rate of technical efficiency change was 0.2–1.2% and the contribution rate of technical change was 0.6–0.8%. Based on the two-stage data envelopment analysis (DEA) method, this study analyzed the operational efficiency of fish farming in EU member countries and discovered that the productivity of freshwater finfish decreased to 0.802 during the study period. Furthermore, the productivity of marine finfish increased to 0.885, and the average technical efficiency of freshwater finfish was 0.918. A meta-frontier DEA approach was used to compare the efficiency of 625 farms producing pangas and tilapia in different regions and production environments in Bangladesh. The study revealed that the production efficiency of pangas in central-eastern Bangladesh was the highest, while that of tilapia in the northeast and central-east was the highest [27].

In summary, the following factors furnish this study with innovation: firstly, environmental factors are included in the analysis of comprehensive aquaculture production capacity, reflecting the ecological efficiency of aquaculture in China. Secondly, the combination of freshwater aquaculture and mariculture reflects the level of ecological efficiency of aquaculture more comprehensively in China. Thirdly, the Zofio decomposition of the ML index is included to further investigate the change in the green total factor productivity of aquaculture.

The structure of this paper is as follows: the Section 1 is the introduction, which presents the realistic background and research significance of this study; the Section 2 is the literature review, which analyzes relevant research and theories; the Section 3 introduces the empirical method of this study; the Section 4 illustrates the data sources and related indicators of this study; the Section 5 discloses the empirical results and analysis; and the Section 6 discusses forward-looking policy recommendations on aquaculture according to the results.

3. Methodology

Conventional DEA models are usually radial and angular. When there is excessive input and insufficient output, this radial DEA model will overestimate the efficiency value of DMU; while using angle DEA to measure DMU efficiency, the deviation of angle selection cannot be avoided, which will cause the calculated results to be inconsistent with the actual situation [28,29]. To overcome the shortcomings of the conventional DEA model, Tone (2001) proposed an efficiency measurement method based on slack variables, which is the non-radial and non-angle SBM distance function model [30]. The mathematical expression is as follows:

$$\begin{array}{c}{\rho }^{\ast }=\min \left[\left(1-\frac{1}{N}{\displaystyle \sum _{n=1}^N\frac{S_n^x}{x_{k^{\prime }n}^t}}\right)/\left(1+\frac{1}{M}{\displaystyle \sum _{m=1}^M\frac{S_n^y}{y_{k^{\prime }n}^t}}\right)\right]\\ s.t.{\displaystyle \sum _{k=1}^K{\lambda }_k^t}{x}_{kn}^t+S_n^x=x_{k^{\prime }n}^t,n=1,\dots ,N\\ {\displaystyle \sum _{k=1}^K{\lambda }_k^t}{y}_{km}^t-S_m^y=y_{k^{\prime }m}^t,m=1,\dots ,M\\ {\lambda }_k^t\ge 0,S_n^x\ge 0,S_m^y\ge 0,k=1,\dots ,J\end{array}$$

(1)

Equation (1) is a non-parametric linear minimum programming model used to determine the relative efficiencies of similar decision-making units based on identical categories of input and output variables improved by Tone (2003). In Equation (1), $x_{k^{\prime }n}^t$ and $y_{k^{\prime }m}^t$ are, respectively, the input and output values of the kth decision-making unit in period t; $s_n^x$ and $s_m^y$ are, respectively, the input and output slack variables; and ${\lambda }_k^t$ is the weighted coefficient vector. Although the SBM distance function model optimizes the shortcomings of the conventional DEA model, the output of the production process is not solely economic, and the pollutants produced are not expected. Tone (2004) proposed an SBM model that can manage undesirable outputs based on the SBM distance function model [31]. It assumes that there are n decision-making units, and each decision-making unit contains three elements of input, output, and undesirable outputs. The mathematical expression is as follows:

$$\begin{array}{c}\overrightarrow{S^{t+1}(x_{k^{\prime }}^{t+1},}{y}_{k^{\prime }}^{t+1},z_{k^{\prime }}^{t+1})={\rho }^{\ast }=\min \left\{\left(1-\frac{1}{N}{\displaystyle \sum _{n=1}^N\frac{s_n^x}{x_{k^{\prime }n}^{t+1}}}\right)/\left[1+\frac{1}{M+J}({\displaystyle \sum _{m=1}^M\frac{s_m^y}{y_{k^{\prime }m}^{t+1}}}+{\displaystyle \sum _{j=1}^J\frac{s_j^z}{z_{k^{\prime }j}^{t+1}}})\right]\right\}\\ s.t.{\displaystyle \sum _{k=1}^K{\lambda }_k^{t+1}}{x}_{kn}^{t+1}+s_n^x=x_{k^{\prime }n}^{t+1},n=1,\dots ,N;\\ {\displaystyle \sum _{k=1}^K{\lambda }_k^{t+1}}{y}_{km}^{t+1}-s_m^y=y_{k^{\prime }m}^{t+1},m=1,\dots ,M;\\ {\displaystyle \sum _{k=1}^K{\lambda }_k^{t+1}}{z}_{kj}^{t+1}+s_j^z=z_{k^{\prime }j}^{t+1},j=1,\dots ,J;\\ {\lambda }_k^{t+1}\ge 0,s_n^x\ge 0,s_m^y\ge 0,s_j^z\ge 0,k=1,\dots ,K\end{array}$$

(2)

Based on the ML productivity index of the traditional directional distance function, the SBM-ML productivity index based on the SBM distance function from period t to period (t + 1) is constructed. The mathematical expressions are as follows:

$$\begin{array}{ll}SBM\_M{L}_t^{t+1}& ={\left[\frac{\overrightarrow{S^t(x^{t+1},}{y}^{t+1},z^{t+1};y^{t+1},-z^{t+1})}{\overrightarrow{S^t(x^t,}{y}^t,z^t;y^t,-z^t)}\times \frac{\overrightarrow{S^{t+1}(x^{t+1},}{y}^{t+1},z^{t+1};y^{t+1},-z^{t+1})}{\overrightarrow{S^{t+1}(x^t,}{y}^t,z^t;y^t,-z^t)}\right]}^{\frac{1}{2}}\\ & =\underset{S\_MLTC}{\underbracket{{\left[\frac{\overrightarrow{S^t(x^t,}{y}^t,z^t;y^t,-z^t)}{\overrightarrow{S^{t+1}(x^t,}{y}^t,z^t;y^t,-z^t)}\times \frac{\overrightarrow{S^t(x^{t+1},}{y}^{t+1},z^{t+1};y^{t+1},-z^{t+1})}{\overrightarrow{S^{t+1}(x^t,}{y}^t,z^t;y^t,-z^t)}\right]}^{\frac{1}{2}}}}\\ & \times \underset{S\_MLEC}{\underbracket{\frac{\overrightarrow{S^{t+1}(x^{t+1},}{y}^{t+1},z^{t+1};y^{t+1},-z^{t+1})}{\overrightarrow{S^t(x^t,}{y}^t,z^t;y^t,-z^t)}}}\\ & =S\_MLT{C}_t^{t+1}\times S\_MLE{C}_t^{t+1}\end{array}$$

(3)

In Equation (3), S_MLTC and S_MLEC represent the technological progress index and the technological efficiency change index based on the SBM distance function, respectively. When SBM_ML, S_MLTC, and S_MLEC > 1, they, respectively, represent the total factor productivity growth, technological progress, and technological efficiency improvement.

$$\begin{array}{ll}{M}_c^{t+1}& =E{C}_c\times T{C}_c=PEC\times PTC\times SEC\times STC\\ & ={\left\{\frac{E_c^t(x^t,y^t)/E_v^t(x^t,y^t)}{E_c^{t+1}(x^t,y^t)/E_v^{t+1}(x^t,y^t)}\times \frac{E_c^t(x^{t+1},y^{t+1})/E_v^t(x^{t+1},y^{t+1})}{E_c^{t+1}(x^{t+1},y^{t+1})/E_v^{t+1}(x^{t+1},y^{t+1})}\right\}}^{\frac{1}{2}}\end{array}$$

(4)

Equation (4) illustrates the Zofio (2007) index decomposition. The PEC is the pure technical efficiency change index; where PEC > 1, the DMU moves to the front, indicating that the efficiency of management and resource allocation is improved. SEC is the scale efficiency change index—SEC > 1 indicates the scale efficiency increase. PTC is the pure technical change index—PTC > 1 indicates the progress of the technical level. STC is the change of scale technology—STC > 1 indicates an increase in the scale effect of technological change [32].

4. Data

Table 1. Input–output index of aquaculture.

Type	Index	Definition	Unit of Measurement
Input	Labor	Fishery practitioners	Person
	Fry	Number of fry	Billion
	Land	Breeding area	Hectare
Output	Positive output	Output value	Ton
	Negative output	Total pollutants	Kilogram

details the input and output indicators for model estimation. The China Fishery Statistical Yearbook, compiled by the China Fisheries Administration, National Fisheries Technology Extension Center, and China Society of Fisheries, systematically contains relevant aquaculture data for China. This study selected sample data from 10 major mariculture provinces and 31 freshwater aquaculture provinces from 2003 to 2020; some missing and unrealistic values are supplemented by the 3-year average method. The pollutant emission coefficient of aquaculture is derived from the Pollution Discharge Coefficient Manual issued by the Ministry of Ecology and the Environment of the People’s Republic of China. The main pollutants in aquaculture are COD, TN, TP, and NH.

The number of input variables and output variables in the SBM model with undesirable outputs will affect the accuracy of the results. According to the cost and characteristics of aquaculture, three input indicators are selected. The expected output is represented by aquaculture production. The unexpected output is the emissions of COD, TN, TP, and NH in the process of aquaculture. In order to prevent inaccurate results caused by excessive undesirable output indicators, a total pollution index is obtained via the summation of pollutants to construct a green total factor productivity index system for aquaculture:

(1)

Labor—mainly consists of the number of workers used in fishery production.

(2)

The quantity of seed—refers to the quantity of marine and freshwater fries.

(3)

Land—refers to the area of mariculture and freshwater aquaculture.

(4)

Positive output—refers to the production of aquatic products.

(5)

Negative output—refers to the total pollutant emissions, including the sum of TN, TP, NH, and COD emitted.

The pollutant calculation formula is (5), and $Q_j$ refers to the jth pollutant emission of aquaculture in a province (unit: tons). $q$ refers to the output of aquatic products in a certain province (unit: tons); $e_j$ refers to the jth pollutant emission coefficient (unit: kg/ton) of aquaculture in a province.

$$Q_j=q\times e_j\times {10}^{-3}$$

(5)

The total pollutant emissions for this study are (6):

$$Total Pollutans = COD+TN+TP+NH$$

(6)

The descriptive statistics of the main variables are illustrated in

Table 2. Descriptive statistics of input and output indicators.

Type	Criterion Layer	Index	Mean	Std. Dev	Max	Min
Mariculture	Input	Labor	349,394.2	212,727.4	1,030,730	2717
		Fry	6189.927	12,463.52	78,398.2	0.5
		Land	194,307.1	212,727.4	942,050	813
	Positive output	Output value	1,611,999	1,580,558	6,417,769	24,295.27
	Negative output	Total pollutants	20,089.04	14,265.55	101.49	56,777.03
		COD	14,751.47	11,490.4	44,610.76	83.41821
		NH	752.5088	975.4925	4904.535	2.025915
		TN	3883.622	4593.937	23,063.43	15.32582
		TP	701.4386	847.168	4130.135	0.716545
Freshwater aquaculture	Input	Labor	438,898.1	459,416.5	1,876,285	40
		Fry	33.46753	122.67	863.3	0.01
		Land	179,791.5	180,916.6	797,575	3
	Positive output	Output value	626,486	767,446.5	3,047,214	28.53014
	Negative output	Total pollutants	18,483.14	30,157.62	143,908.5	0.247555
		COD	15,751.94	27,686.26	134,022.1	0.175105
		NH	468.0329	597.3597	2393.79	0.00756
		TN	1962.22	2255.142	11,170.4	0.058695
		TP	300.946	413.7083	2168.445	0.006195

. From 2003 to 2020, the average labor input extracted from the sample mariculture area was 3,494,400, and the average labor input from the freshwater aquaculture area was 4,389,000. However, the labor input extracted from both freshwater and mariculture areas demonstrated a rapid downward trend, which was closely related to the transfer of agricultural labor in China. The average input and breeding area of fish seedlings extracted from the sample mariculture area were 6189.9 billion and 1943 million hectares, respectively, and from the sample freshwater aquaculture area, they were 33.468 billion and 1.798 million hectares, respectively. The large standard deviation between the input and aquaculture area of fish seedlings in seawater and freshwater aquaculture is due to the large gap in the aquaculture scale between different provinces, while the growth trend is relatively stable within a single province. The expected output is the value of aquaculture production, with an average production of 1611.99 million kg for mariculture and 626.49 million kg for freshwater aquaculture. The unexpected output is the sum of total nitrogen, ammonia nitrogen, and total phosphorus, with an average of 14,265.55 kg for mariculture and an average of 18,483.14 kg for freshwater aquaculture. The emission quantity follows a gradual growing trend in each province.

5. Empirical Results

5.1. Estimation Results

Figure 1 and Figure 2 below illustrate the changes in green total factor productivity of mariculture and freshwater aquaculture in China, respectively. Figure 3 illustrates the gradual trend in the green total factor productivity of mariculture and freshwater aquaculture. Between 2003 and 2020, the total factor productivity of mariculture and freshwater aquaculture in China was 1.06 and 1.05, respectively. Notably, mariculture green total factor productivity fluctuated considerably prior to 2009, with a maximum of 1.192 in 2007 and a minimum of 0.876 in 2008. Presumably, the Chinese government established strict regulations on pollution emissions after the 18th National Congress of the Communist Party of China. Overall, compared with mariculture, the green total factor productivity in freshwater aquaculture changed smoothly, the fluctuations only occurred in 2016 and 2017, with a maximum value of 1.209 in 2017 and a minimum value of 0.94 in 2016.

5.2. Time Evolution Trends

This section analyzes the trend of total factor productivity and the gradual Zofio index decomposition in mariculture and freshwater aquaculture in China. Fare et al. (1994) obtained different efficiency values through VRS and CRS and further compartmentalized EC in FGLR’s decomposition method (1992) into pure technical efficiency change (PEC) and scale efficiency change (SEC). The Zofio decomposition method can be regarded as the further decomposition of technical change (TC) into pure technical change (PTC) and scale technical change (STC) based on the decomposition of Fare et al. (1994).

Figure 4 and Figure 5 below illustrate the time trend of the Zofio index decomposition. From 2003 to 2020, the pure technical efficiency of mariculture changed to 1.022, indicating that the level of management and resource allocation efficiency was improved during this period. The return to scale change was 1.025, indicating that the scale efficiency improved during this period, and the pure technical change was 1.029, indicating that the technical level also improved during this stage. However, the value of scale technical change was 1.01, indicating that the scale effect of technical revolution only improved slightly, while the pure technical change contributed least favorably to the green total factor productivity. The pure technical efficiency change with regard to freshwater aquaculture was 1.025, indicating that the efficiency of the management and resource allocation of freshwater aquaculture improved during this period. Furthermore, the return to scale was 1.006, indicating that the scale efficiency slightly improved during this period. The change in pure technology was 1.044, and the technological level vastly improved in freshwater aquaculture. The change in scale technology was 1.011, indicating that the scale effect of technological change improved, which is similar to mariculture, as pure technological change accounts for the most growth in green total factor productivity.

5.3. Spatial Evolution Trends

This section analyzes the spatial differences between mariculture and freshwater aquaculture and the provincial heterogeneity of the level of green production due to differences in aquaculture scale and resource endowments. Figure 6 below illustrates the green total factor productivity of mariculture within nine provinces and freshwater aquaculture within 31 provinces in China from 2003 to 2020, which varies significantly between provinces. The average green total factor productivity of mariculture showed positive results, except for Shandong. The fastest growth rate was Guangdong Province, with a growth rate of 11.1%, followed by Zhejiang, Tianjin, and Jiangsu with −10%, 8.2%, and 7.1% respectively. The growth rates of green total factor productivity for the remaining three provinces were 1.2%, 4%, and −0.2% within Liaoning, Guangxi, and Shandong, respectively.

The green total factor productivity differences in freshwater aquaculture between provinces are more evident, among which, the top three provinces are Xinjiang, Shanxi, and Shaanxi, with growth rates of 27.3%, 13.9%, and 12.7%, respectively. Furthermore, the majority of the existing studies did not analyze the unrepresentative provinces of Qinghai, Tibet, and Xinjiang due to the lack of water resources. Therefore, this study suggests that although the aquaculture scale is small, the corresponding fine management and technology application will lead to lower pollutant emissions in order to achieve a greener aquaculture mode, which has important significance for developing countries with water shortages and adverse circumstances with regard to developing sustainable aquaculture to solve food security problems.

As shown in Figure 7, freshwater aquaculture is divided into the eastern, central, and western regions. The eastern region includes 12 provinces (Jiangsu, Zhejiang, Shandong, Hebei, Beijing, Tianjin, Liaoning, Fujian, Guangdong, Guangxi, Hainan, Shanghai); the central region includes 9 provinces (Hunan, Hubei, Henan, Shanxi, Jiangxi, Anhui, Inner Mongolia, Heilongjiang, Jilin); and the western region includes 10 provinces (Sichuan, Yunnan, Shaanxi, Gansu, Chongqing, Ningxia, Xinjiang, Tibet, Qinghai, Guizhou). Mariculture is divided between the Yellow Sea, the East China Sea, and the South China Sea areas. The East China Sea includes three provinces (Jiangsu, Zhejiang, Fujian), the South China Sea area includes three provinces (Hainan, Guangdong, Guangxi), and the Yellow Sea area includes four provinces (Shandong, Hebei, Tianjin, Liaoning).

Table 3. MGTFP and Zofio Index decomposition of Mariculture in three regions, 2003–2020.

Region	GTFP	PEC	SEC	PTC	STC
Yellow Sea	1.030	1.012	1.028	1.019	1.005
East China Sea	1.072	1.023	1.015	1.030	1.014
South China Sea	1.056	1.035	1.029	1.041	1.013

and Figure 8 clearly illustrate the green total factor productivity and Zofio index decomposition of the three major regions of mariculture in China. The grey dots in the figure represent the value of each indicator of mariculture in different years. The data revealed that the East China Sea area had the highest green total factor productivity of 1.072, followed by the South China Sea’s factor productivity of 1.056, and the Yellow Sea area had the lowest factor productivity of only 1.030. In the East China Sea area and the South China Sea area, the growth of green total factor productivity mainly depended on pure technological change, with growth rates of 3% and 4.1%, respectively, indicating that technological progress in the region had a considerably positive role in promoting green total factor productivity. In the Yellow Sea area, the change in return to scale had a prominent role in the growth of green total factor productivity, with a growth rate of 2.8%. From the perspective of regional division, the pure technical efficiency change; the change of return to scale; pure technical change; and scale technical change in mariculture areas improved, and the degree of which affected the improvement level of green total factor productivity in each region. In terms of pure technical efficiency changes, the South China Sea area had the highest growth rate of 3.5%, followed by the East China Sea area and the Yellow Sea area, with growth rates of 2.3% and 1.2%, respectively. The degree of improvement with regard to the return to scale of mariculture was the smallest, with the highest rate being located within the South China Sea area with a growth rate of 2.9%, followed by the Yellow Sea area and the East China Sea area, with 2.8% and 1.5%, respectively. The improvement degree of scale technology change was the highest in the East China Sea area, with a growth rate of 1.4%, followed by the South China Sea area and the Yellow Sea area, with growth rates of 1.3% and 0.5%, respectively.

Table 4. FGTFP and Zofio index decomposition of mariculture in three regions, 2003–2020.

Region	GTFP	PEC	SEC	PTC	STC
Eastern Region	1.038	1.020	1.016	1.030	1.007
Central Region	1.046	1.013	0.998	1.044	1.002
Western Region	1.098	1.040	1.001	1.062	1.025

and Figure 9 report green total factor productivity and Zofio index decomposition in the three major regions of China’s freshwater aquaculture. The grey dots in the figure represent the value of each indicator of freshwater aquaculture in different years. Analogous to many studies, the green total factor productivity of freshwater aquaculture in the western region of China was the highest, with 1.098, followed by 1.046 in the central region and 1.038 in the eastern region. The growth of green total factor productivity in freshwater aquaculture is mainly attributed to the improvement of pure technological change, with growth rates of 3%, 4.4%, and 6.2% in the eastern, central, and western regions, respectively. The contribution of pure technical efficiency changes to green total factor productivity cannot be overlooked, with growth rates of 2%, 1.3%, and 4% in the eastern, central, and western regions, respectively. The factors hindering green total factor productivity in various regions are distinct. From the decomposition point of view, the pure technical efficiency, scale efficiency change, pure technical change, and scale technology in the eastern region improved; however, the improvement of scale technology change was not optimal, as this was lower than that of the western region. The pure technical efficiency, pure technical change, and scale technology in the central region improved, but the scale efficiency change was less than one, thus, indicating that the region did not achieve economies of scale in technological innovation. The pure technical efficiency, scale efficiency change, pure technical change, and scale technology in the western region improved; however, the improvement of scale efficiency change was not optimal, thus indicating that the contribution of scale economy in innovation with regard to green total productivity in the western region was not as expected, which is presumably related to the spatial layout of science, technology, and economy in China.

6. Discussion

China’s aquaculture has generated vast economic benefits; however, pollutant emissions cannot be overlooked. Thus, it is necessary to study how to improve the green production efficiency of aquaculture and reduce pollutant emissions. Based on the representative sample data from 10 major mariculture provinces and 31 freshwater aquaculture provinces from 2003 to 2020, this study used the SBM-ML model to calculate the green total factor productivity of aquaculture in China and conducted Zofio index decomposition. This work is of considerable significance with regard to studying the cleaner production of aquaculture in China.

From the perspective of time evolution trends, the green total factor productivity of mariculture in China was 1.050. The decomposition of green total factor productivity demonstrates that pure technical efficiency change, return to scale change, pure technology change, and scale technology change were improved and optimized. Moreover, technological progress had the greatest contribution to increases in GTFP. The green total factor productivity of freshwater aquaculture in China was 1.060. While changes in pure technical efficiency, return to scale, pure technical, and scale technical were greater than one, the growth of green total factor productivity can mainly be attributed to changes in pure technical efficiency and return to scale.

From the perspective of spatial evolution trends, the green total factor productivity of mariculture in the East China Sea area was the highest, followed by the South China Sea area and the Yellow Sea area. The decomposition of green total factor productivity demonstrates that pure technical efficiency change and scale technology change made significant contributions. In the East China Sea area, the active explorations of modern developmental patterns of mariculture have a vital role in optimizing the structure of aquaculture varieties and boosting the transformation and upgrading of modern fisheries, effectively promoting the sustainable development of mariculture. In recent years, organizations such as the China Blue Sustainability Institute (shortened to China Blue) have promoted policy innovation and the implementation of green development policy in the South China Sea area, promoting the high-quality development of aquaculture. Enteromorpha prolifera blooms have serious effects on the level of sustainable mariculture, and, thus, they will have an important impact on the environment in the Yellow Sea area, which may be an important reason why the Yellow Sea region has the lowest green total factor productivity level. The green total factor productivity in the western regions of freshwater aquaculture was the highest, followed by the central and eastern regions. The decomposition of green total factor productivity was evident in the improvement and optimization of pure technical efficiency change, return to scale change, pure technical change, and scale technical change in the western and eastern regions, indicating that all the aspects of green total factor productivity play a crucial role in promoting the green production of aquaculture in China. Moreover, the change in scale efficiency in the central region was less than one, indicating that the region did not achieve economies of scale in innovation. Those spatial differences are strongly related to factors such as the level of economic development, resource endowment, and market demand. The western region is subject to low demand for aquatic products influenced by social factors such as dietary habits; thus, most aquacultural production is on a small-scale with a high level of fine management and green production. The central region features a dense network of rivers and lakes, and the ecological compensation effect of aquaculture has a positive impact on aquaculture. The green production level is lower in the eastern region and is affected by the higher pollutant emissions aroused by its high market demand for aquatic products and the large scale of aquaculture. In summary, technological innovation, industrial spatial layout, input factors, and modern production patterns are important aspects of improving the sustainable development of Chinese aquaculture.

This study produced a relatively comprehensive and systematic analysis of the cleaner production mode of aquaculture in China, which can be used as a reference for future policymaking and future research. The purpose of this study was to highlight the importance of cleaner production with regard to aquaculture and enhance the level of green productivity of aquaculture under the perspective of sustainable development. At any rate, the possibility still remains for further research. Aquaculture, in addition to emissions of pollutants, will absorb part of the pollution emissions. With regard to the availability of data, this study did not consider this effect in the analysis framework.

7. Conclusions and Policy Recommendations

Most previous studies used the DEA model to analyze the production efficiency of aquaculture without considering the environment as an important factor. This study used the SBM-ML model considering undesired output to analyze changes in the green total factor productivity of mariculture and freshwater aquaculture in China, and further evaluated the factors affecting green total factor productivity by employing the decomposition of the Zofio index, which offers a framework for the green productivity of aquaculture. The results indicate that the total factor productivity index of mariculture and freshwater aquaculture in China was 1.050 and 1.060, respectively. Regionally, mariculture in the East China Sea region has the highest green total factor productivity of 1.072, followed by the South China Sea region with 1.056, and the green total factor productivity in the Yellow Sea region is the lowest—only 1.030. The results of the freshwater aquaculture calculations showed the opposite geographical distribution characteristics in China, with the highest in the western region (1.098), followed by the central region (1.046), and the lowest in the eastern region (1.038).

The following policy recommendations based on empirical analysis are provided:

(1) Increasing investment in aquaculture technology. Technological progress has had a significant role in promoting the growth of green total factor productivity. However, in some provinces, pure technological change was not optimized between 2003 and 2020, which hindered the further improvement of green total factor productivity. Government investment has a positive effect on the cultivation and transformation of advanced technologies in aquaculture; thus, sufficient financial support to the development of fishery science and technology should be increasingly provided by the government, consolidating the basic driver for technological innovation in aquaculture.

(2) Optimizing the spatial layout of aquaculture. According to the empirical results, whether it is freshwater aquaculture or mariculture, green total factor productivity had obvious heterogeneity among different regions. To improve the cleaner production level of aquaculture, the efficiency of spatial layout needs to be fully considered in designing support policies for industrial development. For example, freshwater aquaculture should be located in regions with large freshwater areas, such as lakes and reservoirs, so as to increase the scale efficiency and the allocative efficiency for resources, and also to strengthen the role of large-scale aquaculture in water ecological restoration.

(3) Improving the structure of input factors. Aquaculture in China has developed rapidly in recent years with extensive inputs. However, problems, such as weak input structure and low-efficiency management, restrict the high-quality development of aquaculture. Policymakers should focus on optimizing the industry structure to formulate relevant measures to raise the productivity level and management level of fishermen and optimize the overall resource allocation in order to promote the green production of the entire aquaculture. For example, the use of synthetic feed will lead to excessive nitrogen and phosphorus in freshwater, which will negatively affect water quality and ecosystem balance through a process known as eutrophication. The improvement of input structure requires the use of less synthetic feed and the adoption of green natural feed so as to protect the ecological system for aquaculture.

(4) Developing sustainable production patterns. For example, large aquaculture platforms and the deep-water cages, in the process of mariculture, have the advantages of strong resistance to wind and waves, and developing such production patterns can reduce the production risks and contribute to realizing the sustainable development of mariculture. Furthermore, it is critical to take measures to construct an environmental protection system, such as improving manure collection systems and removing illegal fishery cages and farming facilities.