Structural Reform, Technological Progress and Total Factor Productivity in Manufacturing

Abstract

Beneficial institutional environments help to improve TFP in manufacturing. However, it is unclear whether China’s structural reform has led to improvements in TFP in the manufacturing industry. In this study, Sys-GMM was employed to assess the relationships between structural reform, technological progress and manufacturing TFP using panel data from provincial-level administrative regions in China from 2010 to 2020. The results showed the following: (1) Structural reform induced a spiral-shaped upward trend of TFP by affecting the technological progress levels of enterprises in China, which was also enhanced by improvements in regional technology levels. (2) The impact of market-oriented reform on industrial TFP was similar to that of entire structural reform; however, innovation environments did not significantly impact industrial TFP, technological progress had no regulatory effect and although the reform and the supply of public services could significantly promote improvements in industrial TFP, technological progress also exhibited no significant regulatory effect on these improvements. (3) The flow of technology and labor significantly, positively correlated with industrial TFP, while resource idleness had a salient inhibitory effect. In addition, enhancement of the disclosed comparative advantage, as well as further opening and expansion of operation scales of the manufacturing industry, could improve industrial TFP even more.

1. Introduction

The total factor productivity (TFP) of the manufacturing industry of a country or region is an indicator of the economic strength and competitiveness of that country or region. Because of the differences between various TFP estimation methods and gathered data, results can vary considerably from one estimation to another. However, most published studies agree that while the industrial TFP of China has increased, this increase has not been salient, and the industrial TFP level has fluctuated considerably over time [1]. Before 2000, the institutional changes that were brought about by China’s structural reform induced improvements in the technical efficiency of the manufacturing industry, which promoted TFP growth at a relatively rapid rate. After that, the growth rate of TFP followed a continuous downward trend in almost all regions across the country. From 2012 to 2017, the northeast region experienced a negative TFP growth rate [2]. Although the TFP level has stabilized and rebounded in recent years, it remains at a low level. Among the sources of TFP growth, capital investment still ranks first in terms of its contribution to overall industrial growth. This demonstrates that China’s manufacturing industry has not yet entered the stage of efficiency-driven development. Given the complex, changing and uncertain external environments, as well as the new reality facing the domestic economy, it has become even more urgent and important to improve the TFP of China’s manufacturing industry.

Since Krugman questioned the development model of the East Asian economy [3], the study of TFP has become one of the hot topics in many academic circles. Scholars have discussed this topic from a variety of perspectives. At present, research on the impacts of the institutional reform on TFP has attracted considerable attention. A review of relevant papers in the literature showed that the published studies in this field have mainly focused on the following issues.

It is generally assumed that measures for improving market mechanisms (including the promotion of marketization and improvements in laws and regulations) can significantly improve the TFP of enterprises. However, the relationship between the degree of marketization and enterprise TFP is not linear; instead, it is an inverted U-shaped relationship that varies from one ownership type to another. This is largely because resource misallocation affects enterprises with different ownership types differently [4]. The impact of the legal environment on TFP mainly depends on the degree of matching between the way in which the legal systems (such as the environmental regulation system) are set up and the levels of economic and social development. In general, local environmental legislation can significantly improve TFP. However, excessively draconian environmental regulations can incur high costs for enterprises, thus negatively affecting their innovation capacity [5], which is not conducive to improving TFP.

Service-oriented governments reduce institutional transaction costs and governmental interventions and promote firm entry, R&D and innovation, which ultimately leads to an increase in TFP [6]. The reform of the administrative examination and approval systems has had a significant positive impact on the total factor productivity of enterprises in China. Additionally, the intensification of market competition has become an important intermediate channel. The key lies in simplifying the administrative examination and approval processes, which increases the probability of enterprise entry and the entry threat of new enterprises [7]. This positive effect also comes from improving the resource allocation efficiency of enterprises, especially private enterprises, enterprises with high financial constraints and enterprises in regions with poor business environments [8]. However, Li offered the contrary opinion that administrative examination and approval systems significantly inhibited the aggregated productivity growth of China’s manufacturing industry due to the fact that the administrative examination and approval processes distorted the effect of resource allocation and restrained the contribution of the net effect of firm entry and exit to aggregated productivity growth [9].

The expansion of fiscal expenditure for education, public services and social safety has had a positive role in promoting economic TFP and technological progress in China [10]. Wyatt found that both the scale and composition of government spending had significant effects on the level and growth of TFP in OECD countries. Additionally, education spending has played a key role in economic growth, where it is usually assumed to result in the creation of productive human capital [11]. There is also a significant inverted U-shaped relationship between the effectiveness of public transport services and urban green economic growth in China [12]. However, not all research results have supported this view. Nguyen-Van argued that there has been no evidence regarding the impact of national and local public spending on TFP and economic growth in Vietnam’s provinces [13].

The research mentioned above mainly focused on the impacts of institutional environments on industrial TFP from single perspectives. In some cases, although scholars have had different views, they have verified the impacts of market-oriented reforms, government function optimization (with administrative examination and approval systems as the main focus) and public service expenditure on production efficiency. However, studies have rarely analyzed the impact of overall structural reform on China’s manufacturing TFP or the mechanisms through which institutional reform influences industrial TFP, especially the construction of institutional environments that are conducive to innovation. Therefore, to close these research gaps, this study explored the following research questions:

(1) Can structural reform change the industrial TFP of large developing countries during a transition period?

(2) What are the mechanisms of this influence?

(3) How effective are the different contents of structural reform?

To answer these questions, a structural reform indicator scheme for China was constructed from the perspective of improving the innovation capacity of enterprises. The impact of structural reform on China’s manufacturing TFP was assessed from the perspective of promoting technological progress. Based on observations from the existing literature, the marginal contributions of this paper are as follows: (1) an analysis of the mechanisms through which institutional changes drive changes in industrial TFP, which could provide a theoretical basis for subsequent modeling and empirical analysis; (2) the construction of a structural reform indicator system using a comprehensive evaluation method, with the aim of enhancing the innovation capabilities of enterprises; (3) a comprehensive analysis of the effects of structural reform in China and its sub-reforms in different dimensions on changes in industrial TFP.

The remainder of this paper is organized as follows. Section 2 discusses the mechanisms through which institutional changes drive changes in industrial TFP and some of the research assumptions. Our structural reform indicator scheme for China is presented and tested in Section 3. The model settings, variable selection and data description are evaluated in Section 4. Our empirical study and results analysis are presented in Section 5. Ultimately, Section 6 concludes the paper with our conclusions, potential policy implications and further research opportunities.

2. Theoretical Analysis and Research Assumptions

In general, the pace of technological progress, the quality of resource allocation and the magnitude of scale effects are the most direct and important influencing factors of changes in TFP. From a decision-making perspective, these influencing factors are the results of various decisions made by enterprises for the basic goal of profit maximization. Therefore, exploring the forces that affect the profits of enterprises is key to understanding the change mechanisms of industrial TFP. Institutional environments represent unavoidable external conditions for any enterprise during its operation, and the various costs and benefits brought about by institutional settings also constitute decision-making variables for enterprises. A relatively sound production environment is a prerequisite and a key factor for improving industrial TFP [14], and institutions set the basic rules for shaping this kind of environment.

The school of new institutional economics posits that institutions implement rules or forces that people can control and utilize in order to drive individuals to work in a predetermined direction [15]. In reality, these market mechanisms are often imperfect, especially for economies in a transition period, mainly because certain institutional designs are unreasonable, such as market monopolies and segmentation caused by the abuse of administrative power or the underdeveloped rule of law. To promote economic growth, governments have promulgated policies and regulations for institutional innovation to improve the market mechanisms and reduce institutional costs. The aim of these measures is to change the ratio of expected costs to expected benefits in the business activities of enterprises and influence decision-making for resource allocation and the direction of effort, thereby inducing TFP growth. The heterogeneity of the purposes of institutional innovation is a key factor in determining the initial sources and differences in the dominant forces of industrial TFP changes. The structural reform measures that encourage enterprises to engage in more innovative activities (such as rationally dividing governmental functions and market functions, improving the rule-of-law system and reducing the costs of corporate research and development (R&D) investments) allow enterprises to obtain higher returns from innovation achievements. In turn, this strengthens the willingness of enterprises to invest in innovation and continuously improve their technology levels. Therefore, these reform measures play central roles in promoting improvements in industrial TFP. Technological progress causes the production possibility curve of enterprises and industries to expand while the original technical efficiency decreases or even becomes ineffective. This situation provides room for improvement in the optimization of resource allocation. As the combination of production factors shifts these production possibility curves, resources are relocated, production scales expand and resource utilization efficiency improves continuously; ultimately, this is reflected in improvements in TFP. These improvements then enhance the competitiveness of enterprises and increase their profitability, further encouraging enterprises to engage in innovative activities. This cycle is repeated again and again, driving continuous improvements in industrial TFP. Of course, if the institutional environments that result from structural reforms are not effective enough to encourage enterprises to carry out innovation activities, promoting technological progress and improving industrial TFP become more difficult. In short, institutional reforms affect industrial TFP through the intermediate variable of technological progress. If the purpose of an institutional reform is to optimize the allocation of resources (for example, if a government introduces a series of policy measures to control a monopoly or to build a unified market for production factors), the structural reform should remove barriers to, and reduce the costs of, resource flow. This then promotes the flow of ineffective production factors toward places producing higher marginal product values, thus improving the TFP of the entire industry. The reallocation of resources may also lead to the expansion of enterprise scales and increases in industry agglomeration, which enables more enterprises and industries to benefit from the economics of scale, thus improving TFP. However, when technological progress is stagnant, i.e., when the production possibility curves remain unchanged, improvements in TFP also halt as soon as the effects of the resource allocation improvements and the economies of scale on TFP diminish and the production factor combinations approach the boundaries of the production possibility curves. TFP improvements and increases in profits in the early stages not only provide financial guarantees for increasing investments in innovation but may also further stimulate the willingness and motivation of enterprises to participate in innovation activities. Consequently, the outward shift of production possibility curves is promoted, which starts a new round of TFP growth that is driven by technological progress. Of course, if the starting point of institutional designs encourages enterprises to expand their production scales (e.g., increasing land supply and reducing capital prices), the measures taken according to such institutional designs are likely to reduce the prices of production factors and encourage enterprises to invest more resources into increasing the output of existing products and expanding their production scales. When the scales of enterprise expand, the R&D expenses apportioned to each production unit remain relatively low, which may encourage enterprises to increase R&D investment, thus improving TFP. In essence, the expansion of enterprise scales is the result of changes in resource allocation. Therefore, the release of scale effects may also yield the impact of resource reallocation on TFP changes. The results of this impact depend on whether the resource flows induced by institutional changes improve allocation efficiency or create new types of “resource misallocation”. When technological progress is stagnant and pure technical efficiency is effective, changes in TFP grind to a stop as the scale efficiency reaches its optimal level. In other words, the TFP improvement effect brought about by pure scale economies eventually vanishes. However, once new technological progress is achieved, new growth spaces and paths for TFP changes become available.

In summary, regardless of the origins of structural reforms, technological progress achieved by enterprises and industries inevitably leads to changes in resource allocation and economies of scale. This forms a positive feedback effect that results in the self-realization of TFP growth. However, if institutional changes fail to induce improvements in technology levels, their impact on TFP gradually reduces to zero. Technological progress plays a central role in promoting changes in TFP, which is why it is the key factor in realizing continuous TFP growth. In the development stage of efficiency-driven economic growth, the following actions have become important goals for structural reform: enhancing the willingness of enterprises to participate in innovation activities; motivating enterprises to increase investments in innovation; promoting technological progress in industries; improving technological efficiency; and, ultimately, improving TFP and industrial development quality. In addition, the flow and reallocation of resources are also indispensable conditions for changes in TFP. Therefore, we proposed the following hypotheses:

Hypothesis 1 (H1).

Structural reform induced a spiral-shaped upward trend of TFP by affecting the technological progress levels of enterprises, and the technological progress had a regulatory effect.

Hypothesis 2 (H2).

Industrial TFP can be improved by optimizing resource allocation and implementing economies of scale.

The core notions of structural reform in China are to give play to the roles of both the government and the market, ensure that the market plays a decisive role in the allocation of resources, create favorable environments for innovation, encourage enterprises to increase their innovation efforts and achieve sustainable development. In addition, some public services, such as social security and environmental protection, are also essential elements in the process of economic development and can affect the innovation activities of enterprises. In other words, the promotion of marketization, the construction of innovative environments and the enhancement of public services constitute different dimensions of China’s economic system reform at the present stage. These dimensions affect the innovation investments and abilities of enterprises from different aspects so as to improve the overall enterprise TFP. Improving and reforming marketization further stimulates the endogenous motivation of enterprises to participate in innovation activities from the micro perspective, while the construction of innovation environments centers on directly influencing the innovation behavior of enterprises, and improvements to public services focus more on influencing the innovation activities of enterprises via the spillover effect. Although structural reforms are heterogeneous, they can cause changes in the various costs and benefits of commercial activities, thus affecting the decision-making processes of enterprises. These decisions encourage enterprises to engage in more innovative activities, thereby promoting technological progress, which is accompanied by the optimization of resource allocation and even an increase in scale efficiency. This eventually leads to changes in the TFP of enterprises and industries. Based on the above analysis, we proposed the following third hypothesis:

Hypothesis 3 (H3).

Changes in the TFP of the manufacturing industry in China that were affected by structural reform were related to its heterogeneity.

3. Construction of the Structural Reform Evaluation Indicator System and Value Determination of Indicators

3.1. Principles for the Construction of the Indicator System

Structural reform indicators should be designed carefully to ensure that the indicators are scientific, systematic and available. Based on the premise of obeying the general laws of the development of socialist market economies, the construction of a structural reform indicator system should be conducive to building innovative societies and realizing the sustainable and high-quality development of the manufacturing industry. Therefore, the following issues should receive particular attention: (1) The system should help to advance the marketization process and clarify the boundaries between the government and the market. Improving the socialist market economy is not only the main direction of structural reform in China but it is also the environmental foundation for improving the TFP of enterprises. (2) The system should emphasize the institutional construction of innovation environments, as enhancing the innovation capabilities of enterprises and improving the quality of their development have become the main themes and urgent requirements of the development of China’s manufacturing industry, and sound institutional environments can increase the willingness of entrepreneurs to carry out innovation activities and improve their innovation capacity, thus providing the necessary external conditions for high-quality economic development. (3) The system should highlight the government’s capacity to provide public services. Under market economy systems, improving public service systems and providing high-quality public services can improve innovation environments and reduce innovation costs. Consequently, it becomes easier for enterprises to engage in innovation activities, which is one of the basic functions of modern governments.

3.2. Indicator System Design

Based on the above analysis, we proposed the construction of a structural reform indicator system for China that would cover the dimensions of market-oriented reform processes, the creation of innovation environments and the optimization of public services. The proposed indicator system comprises 3 level-1 indicators, 7 level-2 indicators, 18 level-3 indicators and 40 level-4 indicators, as shown in Table 1.

Clarifying the boundaries between the government and the market, transforming the functions of the government and strengthening the rule of law are key tasks for improving market economy systems. Based on this perspective, the government–market relationship and the construction of legal environments are treated as the basic components of market-oriented reforms. In this context, our discussion of the government–market relationship drew on the research of Wang et al. [16]. The construction of a rule-of-law society relies on the conscious implementation of the rule of law, beneficial public security environments and legal services, which also constitute the three level-3 indicators of legal environments in our system. Intellectual property protection can provide protection and incentives for corporate innovation activities. Patents can only be obtained by the property rights holder actively applying for them, which reflects the consciousness of the rule of law among residents engaging in technological innovation activities. Additionally, the number of patent applications in a region represents the level of legal awareness in a region [16]. Crime rates [17] and public security expenditure [18] can also be used to measure the legal environments of regions from the perspectives of investments in maintaining the legal environments and the results of these investments. Accessibility to legal services and the convenience of obtaining legal services are important metrics for measuring legal services. In this study, we used the proportion of lawyers and notaries in the population as a measure of legal services [16].

Optimizing innovation environments and providing favorable external environments for innovation activities are important structural reform goals for governments at all levels. Based on the research of Furman [19] and Zhao [20], we constructed an innovation environment indicator system that incorporated three subsets: innovation infrastructures [19], technology service environments [20,21] and business environments [20,21]. The innovation infrastructure subset consisted of three level-3 indicators: an infrastructure indicator [19,21], an economic development indicator [22] and an education development indicator [19,20]. The infrastructure indicator was mainly used to measure the “hardware infrastructures” for innovation, while the economic development indicator and education development indicator were used to measure the “software infrastructures” for innovation. Using the research method that most scholars have employed, internet penetration rate [23] and mobile phone penetration rate [24] were selected as metrics for measuring digital infrastructures. Sound transportation infrastructures are necessary for the development of the manufacturing sector, which affects the level of innovation costs. In this study, the ratio of the sum of railway mileage and highway mileage to the area of a region was used as a measure of infrastructure status [21]. In general, the overall level of economic development reflects the scale and vitality of regional economic development, which influences the innovation performance of enterprises. In addition, the number of high-tech enterprises indicates the innovation capacity of manufacturing enterprises in different regions. Therefore, consumption level per capita [21], GDP per capita [22] and the development levels of high-tech enterprises [21] were selected to estimate the value of the economic development indicator. An inclusive and open social environment is an indispensable environmental prerequisite for innovation activities, and the level of education development can reflect the quality of citizens within a society. Thus, this study used two level-4 indicators, namely education expenditure per capita and the proportion of personnel with college-level education or above among the total population [20], to determine the value of the education development indicator.

Table 1. The evaluation indicator system for China’s structural reform.

Level-1 Indicators	Level-2 Indicators	Level-3 Indicators	Level-4 Indicators
Market-oriented reform processes	Government–market relationship [16]	Resource allocation [16]	The proportion of economic resources allocated by the market [16]
		Government scales [16]	The proportion of government expenditure to GDP [16]
		Market interventions [16]	The scale of government intervention in business [16]
	Legal environments [16]	Legal consciousness [16]	The number of patent applications per capita [16]
		Public security [17]	Crime rates [17]
			The proportion of expenditures for public security within the GDP [18]
		Legal services [16]	The proportion of lawyers and notaries within the population [16]
Creation of innovation environments	Innovation infrastructures [19]	Infrastructure [19,21]	Internet penetration rate [23]
			Mobile phone penetration rate [24]
			Road and railway mileage per square kilometer [21]
		Economic development [22]	Consumption expenditure per capita [21]
			GDP per capita [22]
			The number of high-tech enterprises [21]
		Education development [19,20]	Education expenditure per capita [20]
			The proportion of personnel with a college-level education or above within the total population [20]
	Technology service environments [20,21]	Technology service industrial development [21,22]	The proportion of scientific research and technical service employees in the services sector [22]
			The proportion of value added from scientific research and technical services in the services sector [22]
			The total assets of scientific research and technical services [21]
		Incubator development [21]	The number of technology enterprise incubators [21]
			The number of graduates from technology enterprise incubators [21]
		Technology service financial development [19,21]	The amount of loans obtained from financial institutions [21]
			The amount of venture capital obtained by technology enterprise incubators [19]
			The total amount of the incubation funds of incubators [21]
	Business environments [20,21]	Fair competition [25]	The proportion of net financial expenses of non-state-owned enterprises to total liabilities [25]
			The proportion of sales and income tax paid by non-state-owned enterprises to sales [25]
			The proportion of non-state-owned enterprises within the total number [25]
		Economic freedom [26,27]	The proportion of the total value of import and export goods in GDP [26]
			The proportion of the total value of domestic trade in GDP [27]
Optimization of public services	Social and cultural development [21,28,29]	Social security [30]	The proportion of social security expenditures in fiscal expenditures
			The number of health technical personnel in healthcare institutions per 1000 persons
			The coverage rate of unemployment insurance
		Cultural construction [21,31]	The culture construction costs per capita [21,31]
			The number of public libraries per 10,000 persons [31]
			The number of people in domestic audiences for cultural and artistic performances per 10,000 persons [21]
	Ecological protection [32,33]	Environmental protection [32]	The chemical oxygen demand of wastewater per CNY 100 million of GDP [32]
			The nitrogen oxide emissions per CNY 100 million of GDP [32]
			The investments in industrial pollution control [32]
		Sanitary environments [33]	The harmless treatment rate of urban domestic waste [33]
			Urban green space per capita [33]
			The number of public toilets per 10,000 persons [33]

Table 1. The evaluation indicator system for China’s structural reform.

Level-1 Indicators	Level-2 Indicators	Level-3 Indicators	Level-4 Indicators
Market-oriented reform processes	Government–market relationship [16]	Resource allocation [16]	The proportion of economic resources allocated by the market [16]
		Government scales [16]	The proportion of government expenditure to GDP [16]
		Market interventions [16]	The scale of government intervention in business [16]
	Legal environments [16]	Legal consciousness [16]	The number of patent applications per capita [16]
		Public security [17]	Crime rates [17]
			The proportion of expenditures for public security within the GDP [18]
		Legal services [16]	The proportion of lawyers and notaries within the population [16]
Creation of innovation environments	Innovation infrastructures [19]	Infrastructure [19,21]	Internet penetration rate [23]
			Mobile phone penetration rate [24]
			Road and railway mileage per square kilometer [21]
		Economic development [22]	Consumption expenditure per capita [21]
			GDP per capita [22]
			The number of high-tech enterprises [21]
		Education development [19,20]	Education expenditure per capita [20]
			The proportion of personnel with a college-level education or above within the total population [20]
	Technology service environments [20,21]	Technology service industrial development [21,22]	The proportion of scientific research and technical service employees in the services sector [22]
			The proportion of value added from scientific research and technical services in the services sector [22]
			The total assets of scientific research and technical services [21]
		Incubator development [21]	The number of technology enterprise incubators [21]
			The number of graduates from technology enterprise incubators [21]
		Technology service financial development [19,21]	The amount of loans obtained from financial institutions [21]
			The amount of venture capital obtained by technology enterprise incubators [19]
			The total amount of the incubation funds of incubators [21]
	Business environments [20,21]	Fair competition [25]	The proportion of net financial expenses of non-state-owned enterprises to total liabilities [25]
			The proportion of sales and income tax paid by non-state-owned enterprises to sales [25]
			The proportion of non-state-owned enterprises within the total number [25]
		Economic freedom [26,27]	The proportion of the total value of import and export goods in GDP [26]
			The proportion of the total value of domestic trade in GDP [27]
Optimization of public services	Social and cultural development [21,28,29]	Social security [30]	The proportion of social security expenditures in fiscal expenditures
			The number of health technical personnel in healthcare institutions per 1000 persons
			The coverage rate of unemployment insurance
		Cultural construction [21,31]	The culture construction costs per capita [21,31]
			The number of public libraries per 10,000 persons [31]
			The number of people in domestic audiences for cultural and artistic performances per 10,000 persons [21]
	Ecological protection [32,33]	Environmental protection [32]	The chemical oxygen demand of wastewater per CNY 100 million of GDP [32]
			The nitrogen oxide emissions per CNY 100 million of GDP [32]
			The investments in industrial pollution control [32]
		Sanitary environments [33]	The harmless treatment rate of urban domestic waste [33]
			Urban green space per capita [33]
			The number of public toilets per 10,000 persons [33]

The technology service subset included three level-3 indicators: technology service industrial development [21,22], incubator development [21] and technology service financial development [19,21]. The proportion of employees and the added value of science and technology services in the service industry [22] and the total assets of scientific research and technical services [21] were used to estimate the value of the technology service industrial development indicator. Technology enterprise incubators are service organizations that help high-tech enterprises to reduce the costs of innovation and promote the commercialization of their scientific and technological achievements. In this study, two sub-indicators were selected to measure the development level of incubators: the number of technology enterprise incubators and the number of companies that graduated from technology enterprise incubators in a year [21]. Financial services not only boost innovation activities but have also become one of the key determinants of the success or failure of corporate innovations. Three indicators were used to determine the value of the financial development indicator of regional science and technology services: the amount of loans obtained from financial institutions [21], the amount of venture capital obtained by technology enterprise incubators [19] and the total amount of incubation funds of incubators [21].

In general, both the degree of fair competition and economic liberalization constitute the main components of business environments [20,21]. Therefore, these two factors were used for the business environment indicator. Usually, the more sufficient the development of non-state-owned economies, the more intense the competition. Following the example of Gongxiang [25], the proportion of net financial expenses of non-state-owned enterprises out of the total liabilities, and the proportion of sales tax and income tax paid by non-state-owned enterprises out of the sales revenue of non-state-owned industrial enterprises, were selected to describe the degree of fairness in acquiring factors and the degree of competition among non-state-owned industrial enterprises [25]. The market structure was also used to measure the level of competition among enterprises, and the proportion of non-state-owned industrial enterprises out of the total number of industrial enterprises was used as its surrogate variable. Trade status reflects the degree of economic liberalization in a region from the perspective of market transactions. Inspired by the construction of a global economic freedom indicator, we used the proportion of the total value of the import and export of goods [26] and the total value of domestic trade within GDP to construct the economic liberalization indicator [27].

Providing high-quality public services is a basic requirement for improving social market economic systems. From the perspective of improving the innovation capacity of enterprises, we divided public services into the two dimensions of social and cultural development [21,28,29] and ecological protection [32,33]. Social and cultural development was measured using two level-3 indicators, namely, a social security indicator [30] and a culture construction indicator [21,31]. Ecological protection was measured using environmental protection [32] and sanitary environment indicators [33]. The social security indicator comprised the following three indicators: the proportion of social security expenditure within fiscal expenditure; the number of technical healthcare personnel in healthcare institutions per 1000 persons; and the coverage rate of unemployment insurance. The level of cultural construction was measured via the following three indicators: cultural construction costs per capita [21,31]; the number of public libraries per 10,000 persons [31]; and the number of people in domestic audiences for cultural and artistic performances per 10,000 persons [21]. The environmental protection indicator used the following three sub-indicators: the chemical oxygen demand of wastewater per CNY 100 million of GDP; the nitrogen oxide emissions per CNY 100 million of GDP; and the amount of investment in industrial pollution control [32]. The harmless treatment rate of urban domestic waste, the urban green space per capita and the number of public toilets per 10,000 persons were used to determine the value of the sanitary environments indicator [33].

3.3. Value Determination of the Structural Reform Indicators

Our structural reform indicator scheme consisted of indicators of four levels, in which level-4 indicators were basic indicators. The government–market relationship factor for the market-oriented reform indicator was borrowed from the research of Xiaolu [16]. The original data for all of the other basic indicators were obtained from various statistical yearbooks or were calculated based on data from those yearbooks. As different indicators had different meanings and the statistical data had different dimensions, the data were incomparable. Therefore, it was necessary to perform dimensionless processing on all indicators. In this study, we adopted the dispersion standardization method to process the indicator data. The method was as follows:

$$V_i=\frac{v_i-v_{\min }}{v_{\max }-v_{\min }}$$

(1)

where $V$ denotes the indicator value and $v_i$, $v_{\min }$ and $v_{\max }$ denote the original data, the minimum value and the maximum value of the i-th indicator of the selected sample, respectively.

In our indicator system, seven level-4 indicators (namely, the proportion of resources allocated by the market, the proportion of government expenditure within GDP, the amount of government intervention in business, the proportion of public security expenditure within GDP, crime rates, the chemical oxygen demand of wastewater per CNY 100 million of GDP and the nitrogen oxide emissions per CNY 100 million of GDP) were “reverse indicators”. This means that the larger the original data, the smaller the value of the corresponding indicator and the worse the reform effect. The rest of the indicators were positive indicators. The reverse indicators were reversed using the transformation threshold method, as follows:

$$V_i=\frac{v_{\max }-v_i}{v_{\max }-v_{\min }}$$

(2)

The linear weighting method, which is commonly used in most evaluation systems, was adopted to calculate the values of the indicators. The key issue to address in this method was how the weights of indicators could be determined on all levels. Considering that estimation results lose comparability when the difference between weights becomes excessive over time, Wang et al. used the arithmetic average method to estimate marketization indicators for different time series, based on international experience [16]. This method was employed to calculate the indicator values in this study. The results of these calculations for our structural reform system are shown in Appendix A.

4. Model, Variables and Data

4.1. Establishing the Model

Based on the neoclassical economic theory and the findings of published studies within the field, the influences of different factors on TFP were analyzed, and the following basic model was constructed:

$$TF{P}_{i,t}=C+\alpha \ast Mt{i}_{i,t}+A\ast X_{i,t}+u_{i,t}$$

(3)

where $TFP$ denotes the total factor productivity, $Mti$ denotes the manufacturing technology progress indicator, $\alpha$ and $A$ denote the coefficient matrices, $X$ denotes a control variable, $u$ denotes the random error term, $i,t$ denote the region and time, respectively, and $C$ denotes the constant term. It has been pointed out in the existing literature that knowledge transfer, labor flow, idle resources, changes in returns to scale, openness to the outside world and the disclosed comparative advantage of the manufacturing industry are all important factors that affect manufacturing TFP; so, they were selected as the control variables.

This study focused on the impact of China’s structural reform on manufacturing TFP. Our theoretical analysis showed that the structural reform induced a spiral-shaped upward trend of TFP by affecting the technological progress level of enterprises. Therefore, the structural reform indicator and its interaction variable with the technological progress index of the manufacturing industry were introduced into Equation (3). Considering the lagging effect of TFP changes, a lag term was introduced into the explanatory variables to construct the model, as represented by Equation (4):

$$TF{P}_{i,t}=C+{\beta }_1\ast TF{P}_{i,t-1}+{\beta }_2\ast St{r}_{i,t}+{\beta }_3\ast Mi{i}_{i,t}+{\beta }_4\ast St{r}_{i,t}\times Mt{i}_{i,t}+A\ast X_{i,t}+u_{i,t}$$

(4)

where $Str$ denotes the structural reform indicator and ${\beta }_i$ denotes the coefficient. According to Hypothesis 1, ${\beta }_2$, ${\beta }_3$ and ${\beta }_4$ were expected to be positive.

4.2. Selection of Variables

4.2.1. TFP

To avoid bias in the production function settings, a data envelopment analysis (DEA) was employed to estimate the values of industrial TFP in different regions and for different periods. Fixed assets were estimated using the perpetual inventory method [34], and the method devised by Haojie was used to calculate the initial capital stock. Briefly, the initial capital stock was obtained by dividing the actual fixed capital of industrial enterprises in all regions in 1980 by the sum of the depreciation rate and the average investment growth rate from 1981 to 1990 (assuming the same depreciation rate of 10.96% for all provinces). The labor force input was the average number of workers employed by industrial enterprises in all regions at the end of the year. Considering the past adjustments of China’s administrative regions, the data from the Hainan region were separated from those from the Guangdong province before 1988, and the Hainan region was treated as an independent province. Similarly, the data from Chongqing before 1997 were excluded from the statistical scope of Sichuan, as though Chongqing was already a province-level administrative region at that time. The year 1981 served as the base period for the relevant data. The gross industrial output value and the total fixed asset investment of each region were deflated using the gross industrial output value deflator and the GDP deflator.

4.2.2. Technological Progress Indicator

In the manufacturing sector, technological progress is mainly reflected by three aspects: firstly, R&D investment is not only the foundation of innovation in today’s manufacturing industry but is also a guarantee of continuous improvement in technology levels in the future (R&D investment intensity level is defined as the ratio of the R&D expenditure of industrial enterprises in a region to the operating income of industrial enterprises); secondly, in production processes, improvements in technology levels are also reflected in the upgrading of machinery and equipment, i.e., improvements in the level of physical capital (the level of physical capital is represented by the proportion of the expenditure on technological upgrading in the $t$-th period in the $i$-th region to the total industrial output value); finally, the technological progress of enterprises ultimately translates into new products on the market. Therefore, the proportion of new product sales revenue to the operating revenue of industrial enterprises is used as a representation of the commercialization of technological progress. Following the structural reform indexation method, these three sub-indicators were taken as secondary indicators to construct the manufacturing technology progress indicator, denoted as $Mt{i}_{i,t}$.

4.2.3. Other Variables

Drawing upon existing research and the aforementioned theoretical analysis, the other influencing factors were divided into two categories in this study.

The first category contained factors related to resource allocation. In recent years, the impact of resource allocation on TFP has attracted attention from all walks of life. However, it is hard to accurately measure the degree of resource allocation because of the lack of relevant statistical data in China. Inspired by the research of Bai and Zhang [35], in this study, resource allocation status was measured indirectly by introducing certain factors, such as technology flow, labor flow and resource idleness. Among these factors, technology mobility and labor mobility were used to measure the flow of technology and labor between different industries. Technology mobility was defined as the share of the sum of the domestic technology transaction value and the value of technology import contracts of the $i$-th region in the $t$-th period out of the GDP of that region, denoted as $Te{t}_{i,t}$. Labor mobility was represented by the passenger volume of each region, denoted as $La{m}_{i,t}$. The inventory status of the manufacturing industry reflects the efficiency of factor utilization to a certain extent; therefore, in this study, the resource idleness level was defined as the proportion of inventory to the GDP of a given region, denoted as $Id{r}_{i,t}$. Finally, the three variables were indexed separately.

The second category mainly contained factors related to economies of scale, including returns to scale, the degree of openness to the outside world and the disclosed comparative advantage of the manufacturing industry. The average cost coefficient was used as a substitute variable for the returns to scale. The calculation method for this metric involved a two-step process: (1) the total costs of the main industrial business were divided by the main business income to obtain the average costs; and (2) the average costs from other years were compared to the average costs from the base year to obtain the average cost coefficient, which was then standardized and denoted as $Rt{s}_{i,t}$. The year 2010 was taken as the base period. Since its structural reform, China has fully seized the development opportunities from the last round of globalization. By vigorously attracting foreign investment to compensate for the shortage of domestic capital, China has achieved a rapid expansion of its manufacturing scale. In short, foreign direct investment has provided capital and other production factors for the expansion of manufacturing scale to a certain extent. Therefore, the degree of openness to the outside world was used as an indirect reflection of changes in the scale of the manufacturing industry and was defined as the share of the total foreign capital that was actually used in the $i$-th region in the $t$-th period out of the GDP of that region, which was then indexed and denoted as $Op{d}_{i,t}$. In general, the competitive advantage of an industry is closely related to the scale of that industry. Comparative advantage is often used to measure industrial competitiveness in international trade and industrial product trade accounts for a very large share of China’s foreign trade. Therefore, the disclosed comparative advantage was used to measure the scale of China’s manufacturing industry, denoted as $Rca$. Although $Rca$ is usually used to measure the comparative advantage of industrial competitiveness between countries, it is also applicable to different regions in China. For small economies, resource endowment structures are relatively simple, and consequently, there are no clear regional differences in the comparative advantage of specific industries. However, for large economies, such as China, resource endowment structures vary significantly from one region to another. This enables the manufacturing industry in one region to acquire a comparative advantage over other regions that is similar to a comparative advantage between countries. Therefore, $Mrca$ was regarded as a substitute variable to measure the (standardized) comparative advantage of manufacturing industries in different regions of China.

4.3. Data Description

In this study, the set of panel data selected for analysis consisted of data from 30 provinces, autonomous regions and municipalities across China, spanning the period from 2010 to 2020. Considering the particularity of the industrial structure in Tibet, the panel data from Tibet were not included in the sample set. Similarly, the panel data from Hong Kong, Macao and Taiwan were also not considered. The data were obtained from the China Statistical Yearbook, China Science and Technology Statistical Yearbook, China Fixed Asset Investment Statistical Yearbook, China Torch Statistical Yearbook, China Social Statistical Yearbook, China Fiscal Statistical Yearbook, Chinese Lawyer Statistical Yearbook, China Law Yearbook, the statistical yearbooks of provinces, autonomous regions and municipalities and the authoritative data that were published by relevant departments. Because of the lack of fixed asset investment indicators, both the industrial fixed asset investment and the GDP of each region were deflated using the GDP deflator, with 2008 serving as the base period. At the time of writing, the most recent data for the marketization indicator were from 2019. Following the processing method devised by Yu Honghai [36], we took the average growth of the marketization indicators in each region in the period from 2011 to 2019 as the growth in 2020. On this basis, the 2020 marketization indicators for all regions were estimated.

5. Empirical Study

5.1. Endogeneity Problem

It is widely recognized that inverse causal relationships may exist between explanatory variables and explained variables (as exemplified by the fact that improvements in industrial TFP further accelerate the flow of resources and encourage enterprises to increase their R&D investments). This led to endogeneity problems in Equation (3). Using the ordinary least-squares method for estimation inevitably causes large errors in parameter estimation, which could lead to distortions in the explanation of economic phenomena. The methodologies of Lahlou and Navatte, who used the IV-2SLS approach, were followed in this study in order to alleviate the potential endogeneity problem. The lag term is related to the current period but not to the current interference item, which meets the conditions of tool variables. Therefore, we selected the lag periods of each endogenous variable as the tool variables. When the lag term of the explained variables was introduced into the equation as an explanatory variable, a correlation between $TF{P}_{i,t-1}$ and $u_{i,t}$ was inevitable, which raised the endogeneity problem. The GMM system (Sys-GMM) method proposed by Blundell and Bond [37] can effectively solve this problem and was employed in this study.

5.2. Empirical Results

5.2.1. Baseline Results

As presented in the first four columns of

Table 2. The results of our empirical analysis of the impact of China’s structural reform on manufacturing TFP.

Explanatory Variable	TFP
	(1)	(2)	(3)	(4)	(5)	(6)
	OLS	OLS	OLS	IV-2SLS	Sys-GMM	Sys-GMM
$Mt{i}_{i,t}$	0.103 *** (0.000)	0.072 *** (0.000)	0.066 *** (0.000)	0.070 *** (0.000)	0.078 *** (0.000)	0.077 *** (0.000)
$Te{t}_{i,t}$	——	0.051 ** (0.034)	0.042 (0.421)	0.048 * (0.088)	0.055 * (0.055)	0.051 ** (0.034)
$La{m}_{i,t}$	——	0.082 ** (0.042)	0.066 ** (0.018)	0.069 ** (0.037)	0.078 *** (0.000)	0.072 ** (0.042)
$Id{r}_{i,t}$	——	−0.255 * (0.073)	−0.264 *** (0.000)	−0.273 *** (0.007)	−0.166 ** (0.035)	−0.130 *** (0.010)
$Rt{s}_{i,t}$	——	——	0.061 *** (0.000)	0.054 *** (0.000)	0.038 *** (0.000)	0.041 *** (0.000)
$Op{d}_{i,t}$	——	——	0.327 * (0.084)	0.361 ** (0.035)	0.350 * (0.074)	0.334 * (0.084)
$Mrc{a}_{i,t}$	——	——	0.652 *** (0.000)	0.561 *** (0.000)	0.406 *** (0.000)	0.452 *** (0.000)
$TF{P}_{i,t-1}$	——	——	——	——	0.093 * (0.077)	0.075 ** (0.032)
$St{r}_{i,t}$	——	——	——	——	0.103 ** (0.021)	0.185 * (0.094)
$St{r}_{i,t}\times Mti{}_{i,t}$	——	——	——	——	——	0.037 ** (0.043)
$C$	0.618 *** (0.000)	0.664 *** (0.000)	0.631 *** (0.000)	0.674 *** (0.000)	0.802 *** (0.000)	0.781 *** (0.000)
R-squared	0.654	——	——	——	——	——
Hausman Test	——	63.119 *** (0.000)	63.462 *** (0.000)	——	——	——
Kleibergen–Paap rk LM	——	——	——	45.612 *** (0.000)	——	——
First-stage F-value	——	——	——	39.708 *** (0.000)	——	——
AR(1)	——	——	——	——	−2.125 ** (0.019)	−2.368 ** (0.027)
AR(2)	——	——	——	——	−0.537 (0.735)	−0.421 (0.679)
Sargan Test	——	——	——	——	27.629 (0.874)	25.077 (0.862)
Observations	330	330	330	330	330	330

The data in parentheses represent the p-values: ***, ** and * indicate 1%, 5% and 10% p-values, respectively.

, we estimated the specifications in Equation (3). In order to better reveal the relationships between the different influencing factors and manufacturing TFP, different methods were selected to test Equation (3), i.e., no control variables were considered to estimate the results in column (1) and control variables were added step by step in columns (2) and (3), respectively. The standard errors were always robust to heteroscedasticity and were also clustered by province to address the potential serial correlations in the error terms for particular provinces. The results of the Hausman test showed that it was appropriate to test with a fixed effect. Considering the endogeneity of the variables in Equation (3), the IV-2SLS approach was employed to obtain the results in column (4). In order to ensure the effectiveness of the tool variables in this approach, the Kleibergen–Paap rk LM statistic was used for non-recognizable tests, and weak tool variables were tested using the F-value of the regression results from the first stage. The results indicated that the original assumption was rejected at a 1% level, indicating that the tool variables were appropriate.

The first four columns of Table 2 show that the coefficients of technological progress were greater than zero and were significant at the 1% level, as expected. Note that compared to the results column (1), the coefficient values of technological progress in columns (2) and (3) became significantly smaller after adding control variables, and there is no doubt that explanatory variables were omitted to obtain the results in column (1). After considering the endogeneity problem in column (4), the coefficients increased slightly; however, the significance level of the coefficient of each variable did not change significantly. This means that there was a positive correlation between technological progress and manufacturing TFP, regardless of whether the control variables and the endogeneity of each variable were considered, which was consistent with the theory. However, the huge inertia of the manufacturing development model and the possible role of institutional change were ignored in Equation (3), which could have distorted the impact of technological progress and caused bias in the estimated coefficients. For this purpose, Equation (4) needed to be examined.

5.2.2. Analysis of the Regulatory Effect

To ensure that the GMM system held and that the instrumental variables were effective, it was necessary to perform an autocorrelation test on the random disturbance term and overidentification tests on the instrumental variables, as shown in Table 2. The results suggested that the second-order autocorrelation coefficient of the disturbance term was not significant at the 10% level. This means that the disturbance term was free from second-order autocorrelation, i.e., the GMM system was applicable. The results from the Sargan test showed that the instrumental variables in the model were effective at the 10% level. In short, the use of the GMM system for the estimation of Equation (4) was appropriate.

The results in columns (5) and (6) in Table 2 were based on Equation (3) and included lag terms for manufacturing TFP, structural reform and its interaction index with technological progress. The results showed that the regression coefficients of these three variables and technological progress were all greater than zero and that they were at least significant at the 10% level. Compared to the results for Equation (3), the value of the regression coefficients of each variable fluctuated somewhat; however, the significance level did not change significantly, which not only reflected the omission of explanatory variables from Equation (3) but also indicated the robustness of the model from another aspect.

The results in columns (5) and (6) show that the coefficient of the first-order lag term was greater than zero and that it was significant to at least the 10% level. This means that the change in industrial TFP had a lagging effect. The results in column (6) in Table 2 show that the coefficients of $Str$ and $Mti$ were 0.185 and 0.077, respectively, which were significant to at least the 10% level. More importantly, the regression coefficient of both interaction variables was 0.037, which was significant at the 5% level. According to the regression results, the impact of structural reform and technological progress on industrial TFP could be expressed as

$$\frac{\partial TF{P}_{i,t}}{\partial St{r}_{i,t}}=0.185+0.037Mt{i}_{i,t}^{}$$

(5)

Equation (5) shows that when the technology level of the $i$-th region in the $t$-th period was the lowest in the country, every increase in the structural reform indicator of one unit promoted an increase in industrial TFP by 0.185%. When the technological progress indicator was one unit higher than the minimum level, this yielded an additional growth rate of 0.037% for industrial TFP and vice versa. This means that during the sample period, improvements in the socialist economic system significantly improved industrial TFP. As the manufacturing technology level increased continuously, the promotion effect of the structural reform on industrial TFP was enhanced. This can be explained via the following mechanism: the faster pace of technological progress endows products with higher competitive advantage, resulting in dramatic increases in the profitability of enterprises, hence their increased willingness to engage in innovation activities. At this point, structural reforms provide a good foundation and beneficial external conditions for starting a new round of technological progress, which leads to further improvements in industrial TFP. The law of symmetry indicated that when the pace of structural reform accelerated (especially when it became faster than the lowest level in the country), each increase of one unit helped industrial TFP to increase by an additional 0.037% compared to pure technological progress. This demonstrated that structural reform amplified the improving effect of technological progress on industrial TFP. This is simple to explain: the purpose of a structural reform is to eliminate institutional constraints that hinder innovation in an industry; therefore, the more perfect the system level is, the more conducive it is to carrying out innovation and optimizing resource allocation. At a given technology level, a sound system can help enterprises carry out production with higher efficiency, and new technological progress may even be induced, which further accelerates the growth of TFP. In short, China’s structural reform and technological progress were interconnected, interacted with each other and jointly affected changes in industrial TFP, which verified Hypothesis 1.

Next, the impacts of various resource allocation factors on manufacturing TFP were examined. As shown in column (6) in Table 2, the coefficient values of $Tet$ and $Lam$ were 0.051 and 0.072, respectively, and were significant to at least the 5% level. This demonstrated that accelerated technology flow significantly improved industrial TFP, which was consistent with experience. Additionally, labor flow, which is a form of the reconfiguration of production factors, also helped to improve the TFP of China’s manufacturing industry, which was basically consistent with the prediction of Acemoglu et al. [38]. On the contrary, the coefficient of $Idr$ was −0.130, which was less than zero and was significant to at least the 1% level. This means that resource idleness significantly lowered industrial TFP, which was consistent with reality. This could be because the misallocation of resources caused by resource idleness can reduce the TFP of an entire industry. In fact, if the efficiency of resource allocation among enterprises in China’s manufacturing industry reached the same level as that among US enterprises, China’s industrial TFP would be 30–50% higher than the current level [39].

Finally, the impact of the scale effect on industrial TFP was examined. The coefficient values of $Rts$, $Opd$ and $Mrca$ were 0.041, 0.334 and 0.452, respectively, and they were significant to at least the 10% level. Clearly, the expansion of China’s manufacturing industry has indeed promoted improvements in industrial TFP for a long time. In short, the effects of the disclosed comparative advantage, openness and scale advantage of the manufacturing industry on industrial TFP could be reported in descending order as: disclosed comparative advantage > openness > scale advantage. The increase in the disclosed comparative advantage of the manufacturing industry was indicative of the increase in the export share of China’s industrial products. Industrial products have a greater market than other products, so the expansion of production scale allows enterprises to benefit from the learning effect and scale effect, which is conducive to improving industrial TFP. The higher the degree of openness of a certain country or region to the outside world, the more favorable it is for foreign direct investment. Foreign direct investment can not only help a country to rapidly expand its scale of production, but it can also introduce more advanced production technologies and management knowledge [40]. These positive spillover effects can accelerate improvements in industrial TFP. The change in returns to scale used to be an important source of TFP improvement in China’s manufacturing sector, especially in the early stages of development; however, with the fading of the scale effect, the impact of the expansion of production scale on improvements in industrial TFP has diminished [41]. The diminishing impact of the scale effect, coupled with the ineffectiveness of innovation environments to induce the growth of industrial TFP, resulted in the rapid decline of TFP of China’s manufacturing industry. These results were consistent with Hypothesis 2.

5.2.3. Heterogeneity Analysis of Structural Reform

To examine the impact of China’s structural reform on industrial TFP more closely, the impacts of the three sub-reforms (i.e., market-oriented reform, the creation of innovation environments and the optimization of public services) on TFP were also investigated, denoted as $Go{r}_{i,t}$, $In{e}_{i,t}$ and $Pu{s}_{i,t}$, respectively.

Table 3. The results for the heterogeneity of the structural reform.

Explanatory Variable	TFP
	(1)	(2)	(3)
	Sys-GMM	Sys-GMM	Sys-GMM
$Go{r}_{i,t}$	0.323 * (0.094)	——	——
$In{e}_{i,t}$	——	−0.013 (0.377)	——
$Pu{s}_{i,t}$	——	——	0.008 * (0.057)
$Mt{i}_{i,t}$	0.105 *** (0.008)	0.082 *** (0.000)	0.056 *** (0.000)
$Go{r}_{i,t}\times Mti{}_{i,t}$	0.069 *** (0.000)	——	——
$In{e}_{i,t}\times Mti{}_{i,t}$	——	0.038 (0.275)	——
$Pu{s}_{i,t}\times Mti{}_{i,t}$	——	——	0.005 (0.354)
Observations	330	330	330
AR(1)	−2.571 ** (0.024)	−2.274 ** (0.028)	−2.298 ** (0.017)
AR(2)	−0.571 (0.563)	−0.554 (0.847)	−0.562 (0.582)
Sargan Test	27.571 (0.787)	22.735 (0.874)	25.928 (0.835)

The data in parentheses represent the p-values: ***, ** and * indicate 1%, 5% and 10% p-values, respectively.

shows that the coefficients of the market-oriented reform indicator, technological progress indicator and the interaction index of these two indicators were 0.323, 0.105 and 0.069, respectively. They were significant at the 10% and 1% levels, which was similar to the effect of the structural reform indicator. This means that a reform that focused on promoting marketization could greatly improve industrial TFP. In comparison, the coefficients of the creation of innovation environments, technological progress and the interaction index of these two indicators were −0.013, 0.082 and 0.038, respectively. The regression coefficient of the technological progress indicator was significant at the 1% level, but the other two coefficients were not significant at the 10% level. This indicated that, although technological progress could significantly improve industrial TFP, the creation of innovation environments was not sufficiently effective in improving industrial TFP. This could be because the impact of institutional environments on TFP is affected by both the threshold effect and the marginal diminishing effect. From a different perspective, this also showed that China’s structural reform was relatively slow and that the original development model had considerable inertia. For a long time, expanding the scale of production and maintaining rapid economic growth were the main goals of economic development in China. With the changes in key social problems and the emergence of new problems, governments at all levels have gradually recognized the need to create good innovation environments through structural reforms to stimulate the motivation of enterprises to participate in innovation activities. However, the strong inertia of the crude production model and the relatively slow pace of mentality changes have presented grave challenges for the creation of suitable innovation environments. At present, innovation environments still cannot provide strong support for enhancing the innovation capability of enterprises or improving industrial TFP. In addition, there were no interactions between the creation of innovation environments and technological progress, which could be because the creation of innovation environments could not promote improvements in the technology level of the manufacturing industry and, thus, could not drive changes in TFP.

As shown in column (3) in Table 3, the public service indicator and the technological progress indicator were 0.008 and 0.056, respectively, and were significant at the 10% and 1% levels, respectively. Although the impact of public services on industrial TFP was weaker than that of market-oriented reforms, they could also drive significant improvements in industrial TFP. This could be because social security and cultural construction are conducive to increasing investments in human capital within society as a whole, and environmental protection regulations force enterprises to accelerate their pace of innovation, both of which contribute to improving industrial TFP. However, the regression coefficient of the interaction index between the public service indicator and the technological progress indicator was 0.005, which was not significant at the 10% level. This means that, from a statistical perspective, the impact of public services on industrial TFP was not directly linked to the technology level of the region’s manufacturing industry. In other words, there was no significant mediating effect between the two indicators, which could be determined by the decision-making mechanisms for the supply of public services. In short, there was heterogeneity in the impacts of various components of the structural reform on TFP, as proposed in Hypothesis 3.

5.3. Further Robustness Checks

By examining the relationships between the structural reform, technological progress and industrial TFP, this study enabled us to draw several important conclusions. To verify the reliability of these conclusions, it was necessary to perform a robustness analysis, which still focused on the abovementioned relationships. In view of the close relationship between structural reform and developmental stage, i.e., the structural reform was stage-dependent, the robustness of the surrogate variables was evaluated.

5.3.1. Changing the Value Determination Method for Industrial TFP

To date, researchers have not yet reached a consensus regarding the best method for estimating TFP. Each method that has been utilized so far has had both advantages and disadvantages. In this study, the function method was used to re-estimate TFP for robustness testing. A double-logarithmic production function model was constructed as follows:

$$lnTF{P}_{i,t}=ln{Y}_{i,t}-m_{ii,t}ln{K}_{i,t}-n_{i,t}ln{L}_{i,t}-u_{i,t}$$

(6)

where $Y_{i,t}$, $TF{P}_{i,t}$, $K_{i,t}$ and $L_{i,t}$ denote the total industrial output value, industrial TFP, capital stock and the number of employees in the $i$-th region in the $t$-th year, respectively. The gross industrial production output, capital stock and the number of employees were measured using the method described earlier, based on the results of which the industrial TFP was estimated and the key variables were re-examined. These results are shown in

Table 4. The results of our robustness checks.

	TFP (Changing the Explained Variable)				TFP (Changing the Explanatory Variable)
$St{r}_{i,t}$	0.225 * (0.072)	——	——	——	0.176 ** (0.034)	——	——	——
$Go{r}_{i,t}$	——	0.381 *** (0.000)	——	——	——	0.273 *** (0.000)	——	——
$In{e}_{i,t}$	——	——	−0.045 (0.517)	——	——	——	0.007 (0.194)	——
$Pu{s}_{i,t}$	——	——	——	0.002 ** (0.04)	——	——		0.032 *** (0.000)
$Mt{i}_{i,t}$	0.115 * (0.092)	0.146 ** (0.017)	0.109 *** (0.000)	0.033 *** (0.000)	0.087 ** (0.046)	0.102 ** (0.034)	0.091 *** (0.000)	0.065 * (0.100)
$St{r}_{i,t}\times Mti{}_{i,t}$	0.084 ** (0.026)	——	——	——	0.039 *** (0.000)
$Go{r}_{i,t}\times Mti{}_{i,t}$	——	0.127 *** (0.000)	——	——	——	0.050 * (0.078)
$In{e}_{i,t}\times Mti{}_{i,t}$	——	——	0.074 (0.762)	——	——		0.1324 (0.458)
$Pu{s}_{i,t}\times Mti{}_{i,t}$	——	——	——	0.006 (0.183)	——			0.008 (0.364)
AR(1)	−2.514 *** (0.000)	−2.077 *** (0.000)	−2.836 *** (0.000)	−1.972 *** (0.000)	−2.095 ** (0.044)	−1.937 ** (0.038)	−2.3614 *** (0.000)	−1.271 ** (0.024)
AR(2)	−0.652 (0.516)	−0.308 (0.394)	−0.463 (0.737)	−0.504 (0.582)	−0.892 (0.373)	−0.674 (0.507)	−0.8647 (0.632)	−0.963 (0.244)
Sargan Test	24.462 (0.833)	20.347 (0.672)	30.214 (0.899)	27.013 (0.784)	15.698 (0.906)	17.092 (0.912)	19.3846 (0.897)	22.672 (0.935)

The data in parentheses represent the p-values: ***, ** and * indicate 1%, 5% and 10% p-values, respectively.

. A comparison with the results in Table 2 and Table 3 showed that although the regression coefficients of the key variables had different values and significance levels, they were completely consistent in terms of sign, which verified the reliability of the aforementioned results.

5.3.2. Method to Change the Core Explanatory Variables

In the previous section of this paper, a structural reform indicator system was constructed according to the new tasks that need to be accomplished in the development stage of China’s manufacturing industry. These tasks encompass the sub-reforms of market-oriented reforms, the creation of innovative environments and the optimization of public services. In this study, a structural reform indicator system was constructed based on research of other scholars that was related to the abovementioned three sub-reforms of the structural reform. Drawing on Yang’s concepts [42], the development status of non-state-owned industrial enterprises was used as a substitute variable for marketization. Specifically, the total assets, profits and the number of employees of non-state-owned industrial enterprises were used to measure the level of marketization. The number of patent authorizations per capita was used as a substitute variable for the creation of innovation environments [43]. Public service optimization mainly included quantity and quality optimization, as well as the four sub-indicators of education, medical care, culture and environmental protection, which constituted level-3 indicators [44]. Finally, following the method for calculating the structural reform indicators described above, the indicators were standardized, and weighted calculations were performed. The regression results showed that structural reform promoted improvements in manufacturing technology level, thus improving industrial TFP, which was consistent with existing conclusions.

6. Conclusions, Policy Implications and Further Research

6.1. Conclusions

With the goal of improving the innovation capacity of enterprises, a structural reform indicator system for China was constructed based on the mechanisms of structural reform that have affected changes in industrial TFP. Then, panel data from 30 provinces, autonomous regions and municipalities from the period of 2010 to 2020 were used to verify the relationships between structural reform, technological progress and industrial TFP. The following main conclusions were drawn. (1) Both structural reform and technological progress could significantly improve China’s industrial TFP, and there was a significant interaction effect between these indicators, which showed that structural reform could amplify the improving effect of technological progress on industrial TFP and accelerate the pace of this improvement. In short, the interaction and interdependence of these indicators could drive China’s industrial TFP to grow along a spiral curve, which verified the rationality of the theory. (2) Marketization processes, innovation environments and public services constituted the main components of structural reform, and these indicators had significantly different effects on industrial TFP. The empirical results show that under different sub-reforms of structural reform, technological progress can significantly improve industrial TFP. The impact of marketization on industrial TFP was consistent with the effect of the general indicator of structural reform. However, during the sample period, improving the creation of innovation environments did not effectively improve industrial TFP, and it had no significant interaction with technological progress, i.e., the attempt to create innovation environments by governments at all levels barely promoted the technological progress of the manufacturing industry and, therefore, had almost no impact on industrial TFP. In contrast to marketization and technological progress, improving public services could help to increase industrial TFP significantly, but its interaction index with technological progress was not significant, i.e., improving the supply of public services could promote technological progress and, in turn, increase improvements in industrial TFP, but improving the manufacturing technology level had no impact on the reform of the public service supply system. (3) The flow and reallocation of technology and labor were conducive to improving industrial TFP, whereas resource idleness had a significant inhibitory effect on industrial TFP. This showed that improving resource allocation helped to improve the production efficiency of the manufacturing industry. Enhancing the disclosed comparative advantage of the manufacturing industry, increasing the degree of openness and expanding the manufacturing scale could also improve industrial TFP, but their absolute effects became weaker and weaker. In particular, the expansion of the manufacturing scale had a much weaker impact than the other factors.

6.2. Policy Implications

The policy implications of this paper are clear. Firstly, improving the system could help to improve the TFP of China’s manufacturing industry. Therefore, structural reform must be promoted, the system dividend should be released and enterprises should receive help to improve their innovation capability, all of which could accomplish the transformation of China’s manufacturing industry from big to strong. Regarding the components of structural reform, accelerating improvements in the socialist market economic system could provide more powerful support for improvements in industrial TFP. It is necessary to accelerate the standardization of government functions, build “effective and capable” governments, vigorously promote the rule of law within governments and accelerate the transformation of governmental functions through economic structural reform. These actions could create institutional environments that are beneficial for corporate innovation. Efforts to build more consummate innovation infrastructures, improve the human capital level of an entire society, effectively implement policies that can promote the development of non-state-owned enterprises, create fairer competitive environments and accelerate the pace of improvements in both the creation of innovation environments and ecological construction should be increased. These efforts could effectively motivate enterprises to engage in innovation activities, thus promoting the growth of industrial TFP. Structural reform must be accelerated, and more entities need to be encouraged to participate in the supply of public services as this could improve service quality. Secondly, governments at all levels should ensure the implementation of various policies and measures for the accelerated construction of a large, national, unified market. On the premise of respecting the market, the central government should strengthen regulations and controls to improve the efficiency of resource allocation by promoting the orderly and reasonable flow of production factors. It is important to make full use of both domestic and foreign markets and effectively increase the consumption capacity of domestic residents, thereby laying solid foundations for the expansion of economic scales. At the same time, various measures must be taken to improve China’s degree of openness to the outside world and reduce the institutional costs to be paid by enterprises going global. These measures could provide an external market that improves production efficiency.

6.3. Further Research

This article did not study the relationships between structural reform and resource allocation or the economy of scale, nor did it measure their effects. These aspects should be explored in future research. First of all, it should be verified whether these aspects induce spiral-shaped increases in improvements in the TFP of the Chinese manufacturing industry and whether they have a significant direct connection. Secondly, if there is a significant relationship between them, then the effect sizes between China’s structural reform and technological progress, improvements in resource allocation and economies of scale should be measured to determine which factor has the largest effect so as to provide a reference for the revision of the direction of institutional reform.