The Impact of Technology Innovation on Enterprise Capacity Utilization—Evidence from China’s Yangtze River Economic Belt

Abstract

This paper analyzes the impact of technology innovation on capacity utilization of enterprises located in the Yangtze River Economic Belt through logic reasoning and empirical modeling. Our analysis shows that the mechanism of how technology innovation affects capacity utilization is that the former promotes the latter through meeting market demand, improving production and management efficiency, and optimizing industrial structures. Our empirical results indicate that the influence of technology innovation on the capacity utilization of enterprises in the Yangtze River Economic Belt evidently possesses positive “U” characteristics. Compared with the upstream and downstream regions of the river, the technology innovation of enterprises in the middle reaches can break the U-shaped inflection point earlier. Compared with light industrial enterprises, heavy industrial enterprises can also break the U-shaped inflection point earlier. Compared with non-overcapacity enterprises, those with overcapacity can break the U-shaped inflection point earlier. The technology innovation of non-state-owned enterprises has obvious positive “U” characteristics in the impact of capacity utilization, while the technology innovation of state-owned enterprises has no significant impact on capacity utilization.

1. Introduction

Since the outbreak of the international financial crisis, the global economic downturn has led to shrinking demand, and traditional industries around the world have faced the problem of overcapacity [1]. This problem may increase international trade frictions and cause systemic risks, thereby increasing macroeconomic uncertainty and restricting sustainable economic growth [2]. How to improve capacity utilization (CU) and alleviate overcapacity has become an important topic of concern around the world [3,4]. Among them, China, as a large developing country, faces a particularly serious problem of overcapacity [5]. Since the initial implementation of the reform and opening up policies, China’s economy has maintained high-speed growth. The total economic volume has reached second place in the world, and the scale of the manufacturing industry has expanded rapidly. However, the growth has slowed since the global financial crisis that began in 2008. At the same time, problems, related to economic growth, are gradually emerging. Among them, overcapacity is a prominent problem in the upgrading of China’s industrial structures and the development of improved efficiencies [6,7]. As China’s economy enters a new normal, the pressure to transform and upgrade its industrial structure has increased, the demographic dividend of economic development has disappeared, and structural imbalance has become increasingly serious, resulting in more and more serious overcapacity [8]. Therefore, it is urgent to find effective ways and means to bring the problem of overcapacity under control [9,10]. The cause of overcapacity is rooted in a lack of innovation. Meanwhile, reducing overcapacity is inseparable from improving technology innovation (TI). Through TI, enterprises can continuously improve their own core competitiveness and produce offers that meet the increasingly diversified needs of the market [11]. Therefore, TI has a significant impact on business performance [12] and can be viewed as a point of new economic growth that can help alleviate the excess capacities of enterprises.

China’s innovation capability has obvious regional differences [13]. The Yangtze River Economic Belt is located in the heartland of China and has obvious regional advantages. Since the initial implementation of the reform and opening up policy, the Yangtze River Economic Belt has enjoyed great advantages in terms of both comprehensive strength and strategic support, gradually populating the zone with the highest economic density, except for some coastal areas in China. In March 2014, Premier Li Keqiang proposed for the first time in the government work report that the construction of the Yangtze River Economic Belt should rely on the golden waterway and be treated as a national strategy. The stable development of the Yangtze River Economic Belt is conducive to the stable development of China’s overall economy. Existing researches on this specific region of the Yangtze River Economic Belt mostly focus on environmental carrying capacity [14], carbon emissions [15], green development [16] etc. Although some literatures have also studied the production capacity of the Yangtze River Economic Belt, they focus on energy efficiency [17], and have not directly studied the CU problem in this region. Therefore, focusing on the capacity utilization of enterprises in the Yangtze River Economic Belt has certain theoretical significance for the study of China’s overcapacity problem. To fill this research gap, this paper mainly answers the following questions: what is the impact of technological innovation on capacity utilization? What is the mechanism? Are there firm heterogeneity and regional differences? The marginal contribution of this paper is as follows.

First, this paper provides micro-evidence for the impact of technological innovation on capacity utilization, and expands the relevant research on the influencing factors of capacity utilization. Existing related researches on the impact of capacity utilization mostly focus on government subsidies [7], capital spending [18], tax rate [19], environmental regulation [20] and so on. However, the existing studies rarely discuss the micro-influence of technological innovation on capacity utilization. This paper enriches the micro research on the impact of TI on CU. This paper measures and analyzes the CU of enterprises in the Yangtze River Economic Belt through the data of China’s industrial enterprise database from the perspective of the micro enterprise level. In addition, it studies the impact of TI on the CU of enterprises in the Yangtze River Economic Belt. Our empirical results indicate that the influence of TI on the CU of enterprises in the Yangtze River Economic Belt evidently possesses positive “U” characteristics. Based on our empirical results, this work provides corresponding theoretical guidance for the formulation and implementation of industrial policies in China. Considering the stage of China’s development, the findings of this work are expected to have wide-range reference value for all other developing and underdeveloped countries from around the world.

The second is to analyze the microscopic mechanism of technological innovation’s impact on the CU of enterprises in the Yangtze River Economic Belt. It is found that TI can promote CU by meeting market demands, improving production and management efficiency, and optimizing industrial structures. This paper expands on the existing related research.

Last but not least, based on firm-level and regional-level characteristics, this paper examines the heterogeneous effects of technological innovation on firm capacity utilization, including ownership characteristics, different regions, and whether there is excess capacity. It provides empirical evidence for technological innovation to empower the real economy and lead the high-quality development of enterprises.

The rest of this paper is structured as follows. Section 2 presents the literature review. Section 3 presents the mechanism analysis. Section 4 is the study design. Section 5 discusses the results. The final section yields the main conclusions.

2. Literature Review and Hypothesis Development

TI can affect market demand [21], enterprise production and management [22], and industrial structures [23]. The consequent effect on enterprises’ CU is produced mainly through the following ways:

• Meeting market demand;
• Improving production and management efficiency;
• Optimizing industrial structures.

2.1. Meeting Market Demand

Market demand is the most direct factor to determine the level of enterprise CU [24], representing an equilibrium mechanism of an enterprise’s capacity that automatically coordinates market supply and demand. The main reason for the occurrence of market surplus and the low CU is the backward technology employed and a lack of TI ability of enterprises [25,26,27], which find it difficult to meet the market demand [28]. Therefore, an enterprise with strong innovation ability is more able to use its own products to meet the market demand, and then improve CU.

First of all, TI generally appears on the basis of an enterprise’s ability to analyze market and identify market potential [29], while the level of such an ability is greatly enhanced by the technology innovativeness of the enterprise [25]. When market knowledge is combined with what the enterprise is technological capable of, a timely TI appears. Hence, TI helps an enterprise improve its CU. A potential market demand refers to a demand that is emerging, although it is not yet fully displayed by consumers and cannot be directly sensed by enterprises [21]. On the one hand, enterprises improve their market analysis and research ability based on its level of TI. By fully analyzing the current consumer demand, functionalities, usefulness, and other performance indicators of the available market offers, enterprises can measure the depth and scale of potential demand, and make relatively accurate judgment on the evolution of current market demand. This measurement and judgment provide enterprises with new technology research and development direction and production direction. For example, there is a serious overcapacity problem in China’s coal industry. Increasing TI can improve CU and effectively manage coal overcapacity [30]. Therefore, the improvement in TI can speed up the accurate identification of potential market demand, leading to the creation of products that meet the true needs of the market [31]. That in turn stimulates mass consumption and increase the market purchase volume. This, of course, helps improve the CU rate of enterprises.

Secondly, enterprises upgrade their products through deploying new TI to meet the forever changing market demand. This, of course, improves enterprises’ CU. On the one hand, in the face of continuous demand or critical market demand, enterprises can quickly respond to the changing market demand by constantly carrying out TI, upgrading their existing technology and improving their existing products [32]. Through continuous TI, continuous improvement and innovative design and production of products, and exploration of other creative development paths, enterprises can continuously maintain their competitive advantage and increase their bases of loyal users. So, enterprises must improve their CU in order for them to better meet market demand. On the other hand, by upgrading low-end products through TI, enterprises can reduce low-end supply and increase high-end supply, so as to meet the high-end market demand and improve their CU. Relevant literature research has shown that missing TI is one of the important causes of the low-end product surplus in emerging economics, and also the root cause of the high-end supply shortage in these economies [33]. Through TI, enterprises can realize the development of new products and the transformation and upgrading of their prevalent technologies. Additionally, they can eliminate low-end and backward production capacities and improve high-end supply and demand; they not only avoid the occurrence of “homogenization” product problems by reducing the production of low-end products, but also acquire relatively fast access to the high-end product market through meeting the demand of the high-end market [34]. So, these enterprises’ utilization rate of production capacity can be improved. Therefore, TI can improve CU by meeting market demand. Accordingly, Hypothesis 1 was proposed.

Hypothesis 1.

TI improves CU by meeting market demands.

2.2. Improving Production and Management Efficiency

By capacity utilization, it refers to the ratio of the actual output of an enterprise to its capacity [35]. Therefore, in the case of the same capacity, when the actual output of the enterprise is higher, its CU is also relatively higher. Through TI, enterprises can make their production and management more efficient through effectively coordinating different departments [36]. So, the efficiency of resource utilization is improved, leading to increased output and improved efficiency of CU.

First of all, through TI, enterprises can optimize their management and promote the use efficiency of capital and labor, so they improve their production efficiency and promote the improvement of CU. TI can not only accelerate the transfer of knowledge to the production of consumable products, but also improve labor productivity and make capital investment more efficient in the process of technology search and information exchange [37]. So, TI helps enterprises achieve a constant, or even increased, output under the condition of saving factor investment. This is mainly because enterprises can promote the improvement of workers’ quality and skills through TI [38]. By promoting and applying new technology, enterprises will constantly put forward higher requirements for workers, while providing educational and technical trainings for employees. This, of course, improves their productivities [39]. On the other hand, TI generally makes the original production tools and production processes more efficient. For example, the application of intelligent equipment helps improve not only the quality of an enterprise’ products, but also the production efficiency of the firm [40]. So, TI promotes the improvement of CU.

Secondly, through TI, enterprises can improve their managerial efficiency and optimize the systemic deployment of factors, leading to better utilization rate of capacities. On the one hand, in the process of production, it is necessary to combine land, labor, capital and other production factors in a specific and increasingly better way. Through TI, enterprises can organize and manage these production factors more scientifically, and upgrade the combined effects of the production factors more optimally. Hence, the enterprises’ organizational form and culture, managerial operations and quality can be upgraded [41]. By implementing advanced management concepts and innovatively upgrading prevalent production methods and organizational form and culture, the internal management of an enterprise can become more scientific and efficient. That is to say, TI helps improve the efficiency of enterprises’ managerial efficiency, optimize the more optimized combination of economic elements, and makes enterprises increasingly more efficient in their use of various economic elements and in their cooperation. In the context of increasingly severe environmental regulations, TI has become an important way for enterprises to increase output and improve CU [42,43].

On the other hand, when an enterprise overinvests in one way or another, it will naturally reduce its CU [44]. Through TI, enterprises optimize their management efficiency, inputs of production factors, and CU. For quite a long period of time, Chinese enterprises have adopted the approach of imitating innovation [45]. In terms of independent TI, there is still a large gap between the current Chinese level and the international advanced level. In order to quickly occupy markets while reducing business costs, enterprises tend to introduce complete sets of equipment. Although this approach can temporarily save time and cost, when the same technology is repeatedly introduced into different industrial sectors, similar products will be produced and entrepreneurial competitiveness will be mostly absent. When an enterprise pays attention to its own long-term benefits, it will naturally focus on improving its independent innovation ability. This will gradually help optimize the enterprise’s management, investment in necessary factors and in activities related to TI [46]. Other than relying on the introduction of low-end technology, it will avoid any pursuit of hot spots in low-end markets and reduce the production of similar products that are available in the market. So, TI can further improve the CU. That is to say, TI can improve the efficiency of enterprises’ management, optimize the systemic combination of economic elements and make the cooperation among them more efficient. Accordingly, Hypothesis 2 was proposed.

Hypothesis 2.

TI improves CU by improving production and management efficiency.

2.3. Optimizing Industrial Structure

TI can help transform industrial structures in the direction of high added value [47], and improve the total efficiency of resource use. Hence, it promotes the improvement of the CU of the entire industry.

First of all, TI can regulate the allocation of production factors in the industry, and improve the utilization rate of production capacity. On the one hand, due to differences in technological bases and costs between enterprises and industries, the impact of TI on labor productivity and output efficiency varies from one industry to another, resulting in performance differences. Under the function of price mechanism, all production factors flow spontaneously to sectors with higher productivity [48]. So, enterprises with higher productivity through technology innovation will have more and convenient access to factors of production. On the other hand, an improvement in TI will improve the marginal productivity of each factor and change the relative costs of the factors. So, improving TI strengthens the coupling between economic output structure and factor input structure, promotes the development of the industry, optimizes the industrial structure, and consequently, improves CU.

Secondly, TI drives the formation of new production methods and new industries. Thus, TI improves the utilization rate of production capacity. On the one hand, when the ability of TI in the overall industry continues to improve, it will force the relevant enterprises to carry out corresponding TI activities in order to follow the industrial trend and maintain a competitive advantage. Otherwise, the enterprises will be eliminated by the industrial evolution. Therefore, when each enterprise constantly carries out TI-related activities, the joint effort will further promote industrial development, and drive the emergence of new production methods and new products [49]. The emergence of new products effectively changes the market demand structure, resulting in the consequence that newly emerging products occupy a much wider range of the consumer market than any previously available product. At the same time, backward enterprises will be gradually replaced by enterprises that employ advanced technologies. As a consequence, the development of the entire industry will be promoted so that the industry and enterprise CU will be improved.

Based on the above analysis, we propose the following Hypothesis 3.

Hypothesis 3.

TI improves CU by optimizing industrial structures.

3. Study Design

3.1. The Model

Based on the analysis above of the underlying mechanism and by considering other influencing factors of an enterprise’s CU, this paper establishes the following regression model:

$$\begin{array}{c}C{U}_{i,t}={\partial }_0+{\partial }_{1}Tec{h}_{i,t}+{\partial }_{2}Tec{h}_{i,t}^2+{\partial }_{3}Marke{t}_{i,t}+{\partial }_{4}Financ{e}_{i,t}+{\partial }_{5}Fi{x}_{i,t}+{\partial }_{6}Ope{n}_{i,t}+{\epsilon }_{i,t}\end{array}$$

(1)

where the subscripts $i$ and $t$ represent enterprise and year, respectively, $\epsilon$ is the residual term, $CU$ is the capacity utilization, $Market$ is the market demand, $Finance$ is the financial development, $Fix$ is an enterprise’s investment in fixed assets, and $Open$ is the degree of openness.

3.2. The Variables

3.2.1. The Explained Variable

The explained variable CU (capacity utilization) is as follows: based on the definition of CU based on the cost function method, this paper first deduces the formula of the short-term cost curve of enterprises from the production perspective of the micro-level of enterprises, and then estimates the parameters in the formula through the work of Olley and Pakes [50], in order to solve for the measurement deviation caused by changing factor prices, the single form of cost function and the endogenous problem. The specific methods are described as follows.

First, one must suppose that the capital intensity of each enterprise in an industry segment is the same, while enterprises’ production efficiency is different from one enterprise to another. The production function of enterprises is as follows:

$$Y_{ijt}=A_{ijt}{K}_{ijt}^{\alpha }{L}_{ijt}^{\beta }{V}_{ijt}^{\gamma }$$

(2)

where $Y_{ijt}$, $K_{ijt}$, $L_{ijt}$, $V_{ijt}$ represent the total output, capital stock, labor input and intermediate input, respectively, of enterprise $i$ in industry $j$ in time period $t$. The short-term average cost function of an enterprise can be written as follows:

$$SAT{C}_{ijt}=\frac{1}{Y_{ijt}}\left(P_K^t\overline{K_{ij}}+P_L^t{L}_{ij}+P_V^t{V}_{ij}\right)$$

(3)

where $P$ is the use price of $K$, $L$ and $V$ in each period. Because the capital stock is immutable in the short term, it is used to represent the existing capital stock that cannot be changed in the short term. One must assume that each enterprise allocates resources to maximize its profit. So, we have the following maximization problem:

$$\max {\pi }_{ijt}=P_Y^t{A}_{ijt}{K}_{ijt}^{\alpha }{L}_{ijt}^{\beta }{V}_{ijt}^{\gamma }-P_L^t{\overline{K}}_{ij}+P_L^t{L}_{ijt}+P_V^t{V}_{ijt}$$

(4)

The partial derivatives of L and V in this formula are calculated, and set to equal to 0, by referencing Equations (2) and (3). At the same time, the partial derivatives of (2) are calculated and set to 0. The capacity level with the lowest short-term cost can be solved to obtain the following equation:

$$Y_{ijt}^{*}={(\frac{\beta }{1-\beta -\gamma }\frac{P_L^t}{P_V^t})}^{\beta +\gamma }(\frac{\gamma }{\beta }\frac{P_L^t}{P_V^t})A_{ijt}{\overline{K}}_{ij}^{\alpha +\beta +\gamma }$$

(5)

The CU at the enterprise level is then obtained as follows:

$$C{U}_{ijt}=\frac{Y_{ijt}}{Y_{ijt}^{*}}$$

(6)

In order to avoid the problem of endogeneity, for the parameters in the formula, the method of Olley and Pakes [50] is used in this paper to estimate the production function in Equation (2). In addition, according to the measurement needs, the calculation method of the relevant indicators is described as follows.

First of all, regarding the calculation of capital stock, this paper draws on the method of Jian [51], mainly through the perpetual inventory method, to calculate the capital stock of a company in the current year. The net fixed asset value of an enterprise in 1998 or the inception year is taken as the net asset value of the initial year, converted into the actual value of 1998 according to the capital input price index of each region; and then the investment price index is used to fix the same enterprise for two consecutive years. The difference between the original values of assets is converted into the actual investment value at the price level in 1998; the depreciation value reported by an enterprise is converted to the actual value according to the investment price index; and the capital stock of an enterprise in each year is calculated according to the formula of the perpetual inventory method. Regarding the calculation of the capital input price index, in view of the relatively large gap in fixed asset prices, this paper draws on the method of Jian [51] and calculates the investment price index of each region by deflating capital input.

Secondly, regarding the calculation of labor input, in order to reflect the differences in labor quality among different enterprises, the amount of labor input is divided by the labor price index, and then converted into the actual value of labor input in 1998. As for the calculation of the labor price index, this article first takes the sum of total wages, benefits and labor insurance premiums of the current year as the amount of labor input of an enterprise. Second, this paper divides the calculated labor amount by the corresponding labor employment amount to obtain the labor price of each enterprise.

Finally, this paper compares the labor price in a certain year with the labor price in the previous year to obtain the labor price index for that year. In addition, regarding the calculation of intermediate inputs, this paper directly obtains the intermediate input amount of each enterprise reported in the industrial enterprise database by the corresponding intermediate input price index. As for the intermediate input price index, this article uses the data of Brand [52]. Finally, regarding the output index, it is obtained in this article by dividing the total output value by the output price index. As for the output price index, this article also uses the data of Brand [52].

3.2.2. Core Explained Variable

The core explained variable TECH (TI of an enterprise) is as follows: in relevant literature, its measurement indicators usually include the output and input of TI. Regarding the output of TI, scholars mostly use the number of patent grants [53], the number of new product project developments [54], and the sales of new products [55]. Innovation investment is mainly measured by the R&D density of enterprises. Due to the lack of patent-related data in the database of industrial companies selected in this article, although there are new product output values of various companies, more than 80% of the new product output values are 0, and it is impossible to determine whether it is truly the case or if those companies just did not report the values. Therefore, this paper uses the ratio of R&D expenses of various Chinese enterprises in the database from 2005 to 2007 to the sales of enterprises. That is, this paper uses the R&D input density of enterprises to measure the TI. The larger this ratio is, the higher the TI of an enterprise; and the more importance an enterprise attaches to TI. This article expects that this ratio has a positive impact on CU.

3.2.3. Control Variable

The control variables include market demand, financial development, fixed assets investment, and degree of openness. For recent works on various internal and external determinants of innovation, one can refer to the work of Forrest et al. [56,57]

Market demand (market) can be described as follows: it is the most direct influencing factor that determines the CU of enterprises [24], and a balanced mechanism for the market to automatically coordinate the supply and demand of enterprises’ capacities. When an enterprise faces a mounting magnitude of consumer demand, it naturally has to elevate its CU in order to bring about the desired scale of output. This paper draws on the practice of Braguinsky [24] and adopts (enterprise sales revenue-receivables)/enterprise gross output value to represent market demand.

Financial development (finance) can be described as follows: a good financial system can provide enterprises with information production advantages. Financial institutions can play the role of centralized collection of market information. Such services can help not only reduce the time cost and manpower-related issues for companies to individually collect and analyze market information, but also solve the problem of information asymmetry, while promoting the flow of capital to valuable projects. Thereby, financial development promotes the improvement of enterprises’ CU [58]. This paper uses the ratio of an enterprise’s total debt to its total output to measure the strength of the enterprise’s received financial support. It is expected that financial support has a positive impact on enterprises’ CU.

Fixed asset investment (fix) can be described as follows: enterprises’ fixed capital is generally quite specific and the speed relevant of equipment upgrading is relatively slow. Once the evolution in market demand speeds up sharply, it is prone to accelerated evolution to produce overcapacity [59]. Drawing on the practice of Liu and Sun [60], this paper uses the ratio of the average annual balance of a company’s fixed asset net worth to the company’s total output value to measure the company’s fixed asset investment.es.

Degree of openness (open) can be described as follows: the greater the degree of openness an enterprise has, the more channels the company can use to reach the international market and to expand its sales revenue, and the easier it is for the company to increase its CU [61]. This paper uses the ratio of the export value of an enterprise to the total output value of the enterprise to measure the degree of openness to the outside world.

3.3. Sample and Data Sources

This paper uses Chinese industrial enterprise database data for the timeframe of 1998–2007 to calculate the CU of enterprises in the Yangtze River Economic Belt. The processing of the Chinese industrial enterprise database mainly consists of the following three steps: filter out enterprises in the Yangtze River Economic Belt, construct panel data for the period 1998–2007, and construct relevant indicators.

The step of building panel data includes matching different years, same industry code caliber, and deleting outliers. The China Industrial Enterprise Database provides enterprises’ data by year. Due to the long period of time covered, there is the problem of changing relevant information in the middle of the period. So, it is necessary to build the same logo for the same enterprise in different years. Matching companies in different years basically follows the method of Brandt et al. [52]. In the modification of industry codes, because the National Bureau of Statistics launched a new industry classification standard in 2003, it is necessary to adjust the industry codes before and after 2003 to the same caliber by referring to the method of Brandt et al. [52]. This work uses an approach similar to Li et al. [62] to deal with abnormal values, which includes mainly the deletion of the total industrial output values, net fixed assets, the number of employees, and those intermediate inputs that are less than 0 or missing. Additionally, this research also deletes those enterprises with fewer than 10 employees, original values of fixed assets less than the net values of fixed assets, and gross industrial output values less than the intermediate inputs. By doing so, this article constructs a set of about 250,000 data values of individual entities.

4. Results and Discussion

4.1. Benchmark Regression

In this paper, the fixed effect model (FE), random effect model (RE) and feasible generalized least squares (FGLS) method are used to estimate the model. The regression results are reported in

Table 1. The overall regression results.

Variable	(1) (FE)	(2) (RE)	(3) (FGLS)
Tech	−2.3304 * (−2.22)	−1.6031 ** (−2.49)	−2.0282 *** (−2.89)
Tech2	0.3380 * (1.79)	0.1050 ** (2.02)	0.1277 ** (2.46)
Market	0.0234 (0.28)	0.0152 *** (2.80)	0.0177 *** (3.27)
Finance	0.5002 *** (7.44)	0.6112 *** (11.10)	0.6858 *** (12.12)
Fix	−0.0072 ** (−2.27)	−0.0069 *** (−3.16)	−0.0080 *** (−3.59)
Open	0.0070 (0.14)	0.0068 (0.16)	0.0240 (0.58)
α	0.6499 *** (7.07)	0.6592 *** (17.02)	0.5439 *** (15.05)
Hausman test	0.0048

Note: ***, ** and * represent 1%, 5%, 10% significance levels, respectively, and FE, RE, FGLS represent fixed effects, random effects and feasible generalized least squares, respectively.

.

As can be observed from the regression results in Table 1, the p value of the Hausman test is less than 0.05. So, the null hypothesis is rejected. The fixed effect model (FE) is superior to the random effect model (RE). By comparing the regression results of the fixed effects model and the feasible generalized least square method (FGLS), it can be found that the explanatory variables obtained by the two models have the same signs. Because the FGLS method can eliminate possible heteroscedasticity and sequence correlation to some extent [63], this paper chooses the regression results of Model (3) for analysis.

The regression results of Model (3) show that the first-order coefficient of TI is −2.0282, which passed the 1% significance test; the second-order coefficient is 0.1277, indicating that the impact of TI on the CU of enterprises in the Yangtze River Economic Zone has a significantly positive “U”-shaped feature, with the inflection point at 7.9412. It is shown that at the stage when the level of TI is relatively low, the CU is suppressed, but once the inflection point is passed, TI significantly improves CU. The main reason behind this observation is that in the early stage of TI input, the cost of TI input is relatively high, which creates a squeeze on other production cost inputs, leading to a reduction in CU. In addition, at the initial stage of investment in TI, the transformation of innovation achievements takes a relatively long period of time, without creating any direct market effect quickly. It generally takes a longer time to increase market demand for the consequent market offers and optimize the development of the industry. Therefore, the innovation effect cannot be quickly shown; instead, any potential increase in CU is inhibited. However, when the degree of innovation reaches a certain level, the effect of TI will inevitably be revealed. At this time, enterprises can meet market demands through continuous innovation, optimization of their managerial routines, and reduction in the effects of excessive investment. This will, consequently, significantly increase the utilization rate of production capacity.

4.2. Heterogeneity Analysis

4.2.1. Regional Heterogeneity

The previous section analyzed the overall impact of enterprise CU in the Yangtze River Economic Belt. In the next section, this paper analyzes the possible regional heterogeneity. The regression results are shown in

Table 2. Regional heterogeneity regression results.

Variable	(4) (Upstream Region)	(5) (Midstream Region)	(6) (Downstream Region)
Tech	−2.2914 ** (−1.95)	−4.7877 ** (−2.55)	−3.3203 *** (−2.58)
Tech2	0.1278 * (1.84)	3.5206 * (1.74)	0.5234 ** (2.32)
Market	0.0042 (1.47)	0.1026 (0.54)	0.1267 (1.28)
Finance	0.72668 *** (8.35)	0.3277 *** (4.48)	0.6691 *** (8.87)
Fix	−0.0019 (−1.64)	−0.0516 *** (−6.36)	−0.1488 ** (−10.43)
Open	−0.0400 (−0.22)	0.1418 (1.34)	−0.0574 (−1.18)
α	0.3194 *** (5.92)	0.5218 *** (2.73)	0.5552 *** (5.11)

Note: ***, ** and * represent 1%, 5%, 10% significance levels, respectively.

.

In all regions, the coefficient of first-order TI is significantly negative and the coefficient of second-order TI is significantly positive, indicating that in various regions of the Yangtze River Economic Belt, the impact of enterprises’ TI on CU demonstrates a significantly positive U-type feature. In particular, from the inflection point of the impact of TI of various regions on the promotion of CU, the upstream area is 8.9647, the midstream area is 0.6799, and the downstream area is 3.1718. These points show that the TI of enterprises in the middle reaches break through the U-shaped inflection point earlier than the other areas, with a positive impact on CU. This middle area is followed by the downstream areas, and then the upstream areas. The main reason for this ordering of the regions is that the TI and market demand of the upstream region lag behind those of the downstream region and the midstream region. So, the commercialization of TI results takes a relatively longer period of time, and the effect of TI on CU also takes a longer time to appear. At the same time, its economic level is comparatively more backward, and consumers’ purchasing power is relatively low. When enterprises’ innovation achievements commercially enter the market, consumer acceptance and approval are also relatively limited, resulting in its last-ranked breakthrough in the technological inflection point for upstream enterprises. The reason why the TI of enterprises in the middle reaches breaks through the U-shaped turning point earlier than that of enterprises in the downstream areas is that the TI in the downstream area itself is relatively high, making it relatively more difficult for the area to break through the original level of TI. As for the mid-reach area, by replying to the spillover effect of the TI of the lower reaches, it makes it relatively easier for this area to engage in new TI activities based on the prevalent TI basis. Additionally, its comparatively larger market demand makes it easier for new TI achievements to be commercialized and accepted by consumers. That is why enterprises in the mid-reaches can more easily breakthrough the U-shaped inflection point and improve the CU. This conclusion is similar to the work of Li et al. [64]. Li et al. [64] studied the influencing factors of the green development efficiency of the manufacturing industry in the Yangtze River Economic Belt and found that the R&D investment rate in the middle reaches has the greatest impact on the green development efficiency of the manufacturing industry. Although the research topics are different, they all show that innovation in the middle reaches has brought greater effects to the Yangtze River Economic Belt.

4.2.2. Sector Heterogeneity

In order to explore whether or not the impact of TI on the CU of enterprises in the Yangtze River Economic Zone has industry heterogeneity, this section conducts empirical tests on state-owned enterprises and non-state-owned enterprises, light industrial enterprises and heavy industrial enterprises, surplus enterprises and non-surplus enterprises.

(1)

Heterogeneity of ownership

By comparing the state-owned and non-state-owned enterprises in the Yangtze River Economic Belt (

Table 3. Ownership heterogeneity regression results.

Variable	(7) (State-Owned Enterprises)	(8) (Non-State-Owned Enterprises)
Tech	−0.4192 (−0.09)	−2.0043 *** (−2.80)
Tech2	0.1021 (0.01)	0.1262 ** (2.39)
Market	0.0038 (0.62)	0.0626 (0.55)
Finance	0.4720 ** (2.18)	0.6916 *** (11.95)
Fix	−0.0017 (−0.65)	−0.0097 *** (−3.62)
Open	−0.4264 (−0.88)	0.0054 (0.13)
α	0.1523 (1.13)	0.5155 *** (4.35)

Note: *** and ** represent 1%, 5% significance levels, respectively.

), it is found that among the state-owned enterprises, the coefficient of enterprises’ first-order TI is −0.4192, which is not significant; and the coefficient of second-order TI is 0.1021, which is also not significant. Among the non-state-owned enterprises, the first-order coefficient of enterprises’ TI is −2.0043, which passed the 1% significance test, and the second-order coefficient is 0.1262, which also passes the 1% significance test, showing a significant positive “U” shaped feature, with an inflection point at 7.9409. These observations mean that in state-owned enterprises in the Yangtze River Economic Zone, TI cannot significantly improve the CU, while in non-state-owned enterprises, TI can significantly increase enterprises’ CU. The main reason behind this difference is that the motivation for technological R&D comes from competitive pressures and profit-making desire; but, state-owned enterprises have a natural monopoly position and competitive advantage. Therefore, state-owned enterprises are not driven by technological R&D and do not pay enough attention to TI. Therefore, the chance for state-owned enterprises to experience overcapacity of production is greater than that of non-state-owned enterprises. Lu and Li [65] studied the relationship between the opening of high-speed rail and CU, and discussed the characteristics of enterprise ownership. The study shows that the opening of high-speed rail has a greater positive impact on non-state-owned enterprises CU. Although this paper is different from the research topic, the conclusions on the heterogeneity of enterprise ownership are similar, and they all conclude that the impact on non-state-owned enterprises CU is greater. It shows that non-state-owned enterprises have more motivation to change the capacity utilization rate.

(2)

Differences between enterprises of light and heavy industries

By comparing the enterprises in light and heavy industries in the Yangtze River Economic Belt (

Table 4. Differences between enterprises of light and heavy industries.

Variable	(9) (Light Industrial Enterprise)	(10) (Heavy Industrial Enterprise)
Tech	−3.4846 *** (−3.05)	−3.4323 *** (−3.15)
Tech2	0.2043 *** (2.84)	0.5082 *** (2.56)
Market	0.0155 *** (2.65)	0.0210 *** (2.92)
Finance	0.6547 *** (9.96)	0.6920 *** (9.99)
Fix	−0.0070 *** (−2.85)	−0.0095 ** (−3.09)
Open	0.0887 ** (1.99)	0.0940 (1.63)
α	0.5635 *** (13.35)	0.5588 ** (12.67)

Note: *** and ** represent 1%, 5% significance levels, respectively.

), it is found that for the enterprises in the light industry, the coefficient of first-order TI is −3.4846, and passes the 1% significance test. The coefficient of second-order TI is 0.2043, which also passes the 1% significance test. All of these observations jointly show a clear positive “U” shape characteristic, with an inflection point at 8.5281. For enterprises in the heavy industry in the Yangtze River Economic Belt, the coefficient of first-order TI is −3.4323, which passes the 1% significance test. The coefficient of second-order TI is 0.5082, which also passes the 1% significance test, showing an obvious positive “U”-shaped feature, with an inflection point at 3.3769. Relatively speaking, the TI of heavy-industry enterprises in the Yangtze River Economic Belt break through their U-shaped inflection point earlier than those in the light industry. The main reason behind this phenomenon is that, on the one hand, the production of heavy-industry enterprises is mainly and largely automated, the production process has a high degree of continuity, labor is more finely divided, and although complex, collaborations are better synchronized. Therefore, it is easier for such enterprises to innovate and transform in terms of production equipment and production processes. Hence, they can significantly improve their production efficiency and, thereby, increase the rate of CU. In comparison, for light-industry enterprises, their production processes are mostly completed by human resources. It is relatively more difficult for these enterprises to innovate their production processes and equipment within short periods of time. Therefore, it is more difficult for TI to promote the utilization rate of production capacity of these enterprises. On the other hand, because the products of heavy-industry enterprises are highly targeted, these enterprises design and produce innovative products to directly satisfy market requirements, without much worry about making market prediction errors. On the contrary, the customers and product markets of light-industry enterprises are more complicated, causing difficulty for enterprises to determine the direction of innovation. This difficulty can easily lead to errors in determining innovation directions, which, in turn, can easily cause investment failures, leading to reduction in the CU.

(3)

Heterogeneity of production capacity

Based on a comparison between the excess enterprises and non-excess enterprises in the Yangtze River Economic Belt (

Table 5. Heterogeneity of production capacity.

Variable	(11) (Overcapacity Enterprises)	(12) (Non-Overcapacity Enterprises)
Tech	−2.2926 ** (−1.95)	−2.1247 *** (−2.90)
Tech2	0.1598 *** (1.85)	0.1337 ** (2.48)
Market	0.0042 (1.49)	0.0187 *** (3.05)
Finance	0.7118 *** (8.29)	0.7333 *** (11.25)
Fix	−0.0019 * (−1.67)	−0.0085 *** (−3.29)
Open	0.0420 (0.23)	0.0685 (1.45)
α	0.3288 *** (6.19)	0.5626 *** (13.65)

Note: ***, ** and * represent 1%, 5%, 10% significance levels, respectively.

), it is found that, for excess enterprises Equation (11), the coefficient of the first-order TI is −2.2926, which passes the 1% significance test; the coefficient of the second-order TI is 0.1598, which also passes the 1% significance test, demonstrating an obvious positive “U” shape characteristic, with inflection point at 7.1733. For the heavy-industry enterprises in the Yangtze River Economic Belt Equation (12), the coefficient of first-order TI is −2.1247, which passes the 1% significance test; the coefficient of second-order TI is 0.1337, showing a clear positive “U”-shape characteristic, with an inflection point at 7.9457. Relatively speaking, the TI of excess enterprises in the Yangtze River Economic Belt break through the “U”-shape inflection point earlier, indicating that TI has a certain role in promoting the CU. As for enterprises with excess capacity, most of them belong to traditional manufacturing industries, with their TI held at low levels for a long period of time. Usually, they improve their market size and competitive advantage through expanding investment. Therefore, once these enterprises carry out TI, they will be able to arouse their own advantages of innovation, leading to accompanying resource allocations, increase in market demands, and other positive effects. So, for these enterprises, TI can elevate their CU. Wang and Zheng [25] also studied the relationship between TI and CU, and discussed the characteristics of enterprise capacity. The study shows that TI has a greater positive impact on the CU of non-overcapacity enterprises. For companies with excess capacity, TI cannot significantly promote CU. The conclusion of this paper is not the same. We not only believe that in the long run, TI can significantly increase the CU of overcapacity companies, but also conclude that TIs with overcapacity can break through the “U”-shaped inflection point earlier than non-overcapacity companies.

4.3. Robust Test

This paper performs the robustness test by replacing the explained variable, that is, the calculation method of CU. The Chinese industrial enterprise database does not provide physical indicators for some of the important variables we use instead of value indicators, which include price factors. In this context, the choice of various price indexes may have an important impact on the measurement results. Therefore, this article chooses different price indexes to test the basic results established in the previous sections. By referencing the practice of the “Research on the Policy of Further Mitigating Overcapacity” of the Development Research Center of the State Council, this paper uses each provincial CPI as the labor input deflator of labor price index. By referencing the method of Tang and Wu [66] and by using the intermediate input price index of Brandt et al. [52] to revise the perpetual inventory method of Jian [51], we obtain the deflated capital stock. Lastly, the output price is also deflated by using individual provincial CPIs. All other variables remain unchanged. The regression model still uses the feasible generalized least squares method. The regression results are shown in

Table 6. Robustness test of benchmark regression and regional heterogeneity.

Variable	Overall	Upstream Region	Midstream Region	Downstream Region
Tech	−1.1524 *** (−2.17)	−1.2621 ** (−1.99)	−3.3203 ** (−2.55)	−3.2129 *** (−2.58)
Tech2	0.1345 ** (1.13)	0.1562 * (1.82)	3.5234 * (1.04)	0.2021 ** (2.32)
Market	0.0562 *** (3.44)	0.0123 (1.09)	0.1267 (0.24)	0.1038 (1.28)
Finance	0.6239 *** (4.12)	0.7266 *** (8.93)	0.6619 *** (4.48)	0.4709 *** (8.87)
Fix	−0.8921 *** (−5.34)	−0.0009 (−1.14)	−0.0921 *** (−6.43)	−0.1099 ** (−9.43)
Open	0.0561 (0.13)	−0.02930 (−0.12)	0.1674 (1.18)	−0.5621 (−1.08)
α	0.6499 *** (7.07)	0.2034 *** (7.29)	0.5218 *** (2.73)	0.2341 *** (5.11)

Note: ***, ** and * represent 1%, 5%, 10% significance levels, respectively.

and

Table 7. Robustness test of industry heterogeneity.

Variable	State-Owned Enterprises	Non-State Owned Enterprise	Light Industrial Enterprise	Heavy Industrial Enterprise	Overcapacity Enterprises	Non-Overcapacity Enterprises
Tech	−0.4192 (−0.09)	−2.0923 *** (−2.99)	−3.3910 *** (−2.98)	−3.3463 *** (−2.53)	−2.321 ** (−1.89)	−2.1314 *** (−1.89)
Tech2	0.0209 * (1.09)	0.1091 ** (1.98)	0.1034 *** (2.37)	0.8042 *** (2.04)	0.1391 *** (2.05)	0.1356 ** (2.28)
Market	0.0028 (0.62)	0.0826 (0.35)	0.03915 *** (2.45)	0.0921 *** (2.76)	0.0122 (1.59)	0.0097 *** (3.35)
Finance	0.2309 ** (2.09)	0.4529 *** (9.38)	0.8947 *** (3.42)	0.4090 *** (9.78)	0.5987 *** (7.31 )	0.7333 *** (11.25)
Fix	−0.0021 (−0.35)	−0.0192 *** (−2.89)	−0.0340 *** (−3.09)	−0.0095 ** (−0.23)	−0.0134 * (−1.57 )	−0.0085 *** (−3.29)
Open	−0.3255 (−0.58)	0.0027 (0.83)	0.0497 ** (1.79)	0.0453 (1.83)	0.0292 (0.33)	0.06683 (1.45)
α	0.0981 (1.13)	0.4265 *** (4.05)	0.5424 *** (10.31)	0.4298 ** (10.57)	0.3418 *** (5.45)	0.5324 *** (9.63)

Note: ***, ** and * represent 1%, 5%, 10% significance levels, respectively.

.

By comparing Table 6 and Table 7 and the regression results obtained above, we can observe that the variable signs and relevant significances of the regression results of the overall, individual regions and sub-sectors of the Yangtze River Economic Belt are basically consistent, indicating that the regression results of the overall, individual regions and sub-sectors are relatively robust. That is to say, the impact of TI on enterprises’ CU in the Yangtze River Economic Belt has an obvious positive “U”-shaped characteristic. Through meeting market demands, TI improves enterprises’ managerial efficiency, promotes industrial development, and increases enterprises’ CU.

5. Conclusions

5.1. Main Findings

The findings of this paper show that TI significantly promotes the CU of enterprises in the Yangtze River Economic Belt, and the effect is an evidently positive “U” characteristic. Although this result is similar to those obtained by Wang and Zheng [25], the difference is that Wang and Zheng [25] studied the impact of TI on excess capacity, and concluded that improving competitiveness and reducing investment waste are important ways for TI to reduce excess capacity. The topic of this paper is the CU. This paper explains the mechanism of how TI affects CU from the following three angles: meeting market demand, optimizing an enterprise’s management, and promoting industrial development. Additionally, this paper also empirically tests the promotional effect of TI on enterprises’ CU, and finds that the impact of TI on CU experiences regional heterogeneity and industrial heterogeneity. Although Wang and Zheng [25] also studied the relationship between TI and CU, the conclusions of this paper are not the same. This paper discovers the U-shaped relationship between TI and CU for the first time. In particular, compared with the upstream and downstream regions, enterprises’ TI in the middle reaches can break through the U-shaped inflection point earlier than the other regions. Compared with light-industry enterprises, heavy-industry enterprises can break through the U-shaped inflection point earlier. Compared with non-overcapacity enterprises, those with overcapacity can break through the U-shaped inflection point earlier. The TI of non-state-owned enterprises has an obvious positive “U” characteristic in its impact on CU, while the TI of state-owned enterprises has no significant impact on these enterprises’ CU.

5.2. Policy Implications

Our empirical analysis indicates that the influence of TI on enterprises’ CU in the Yangtze River Economic Belt demonstrates an obviously positive “U” characteristic. Therefore, on the one hand, enterprises should continue to promote their continuous improvement of TI capabilities, thereby improving their CU. Relevant governments should increase investment in scientific and technological innovations, and provide enterprises with support for their independent innovation and R&D through subsidy policies, credit support, and special funds. On the other hand, it is necessary to further strengthen the cooperation between upstream, mid-stream and downstream industries. One must actively build projects of cross-regional industrial cooperation, technology sharing, and collaborative R&D to improve the capacity utilization rate of the Yangtze River Economic Belt. In addition, the management and supervision of SOEs needs to be improved. One must further promote the innovation of state-owned enterprises, forcing state-owned enterprises with backward production capacity to withdraw from the market, truly improve the capacity utilization rate of state-owned enterprises, and control excess capacity.

5.3. Limitations and Future Research

Due to the limitation of data, the conclusion of this paper is based on the sample study of China, which can only provide reference values for other countries, and may not be completely suitable for other countries. In the future, the impact of technological innovation on capacity utilization can be studied based on global data, and heterogeneity across countries can be discussed.