The Financing Efficiency of China’s Industrial Listed Enterprises Based on the Dynamic–Network SBM Model

Abstract

Industry is an important force in China’s economic development; however, with the transformation and upgrading of the industrial structure, a large number of resources have flowed to the tertiary industry, and the funding problem has become one of the main disadvantages restricting China’s industrial enterprises’ sustainable development. This paper aims to point out the problems and improvement directions of financing efficiency of China’s industrial listed enterprises. Based on the two-stage dynamic network SBM (DNSBM) model, this paper evaluates the financing efficiency of 450 of China’s industrial listed enterprises from 2011 to 2017. The results show that: (1) the overall financing efficiency of China’s industrial listed enterprises is low, the funds are not used effectively, and there is great room for improvement; (2) the overall financing efficiency of state-owned enterprises (SOEs) is lower than that of non-state-owned enterprises (NSOEs), the average fund raising efficiency of SOEs is greater than the fund using efficiency, but the opposite is true for NSOEs; (3) the overall financing efficiency of the main-board-listed enterprises is the lowest, and that of the growth enterprise market (GEM) is the highest, the most obvious gap is in the second stage of fund using, but this gap is gradually narrowing; and (4) the overall financing efficiency of China’s industrial enterprises has obvious regional characteristics, the fund raising efficiency value in each region is not much different, while the fund using is significantly different. To improve financing efficiency, enterprises must improve their financing channels, choose the best financing method, maintain a reasonable debt-financing ratio, improve management level and profitability, increase enterprise value, enhance the debt-paying ability, and attract more capital at a low cost. In addition, the government should also provide corresponding financing support policies for different types of enterprises.

1. Introduction

China’s industry has gone through the industrialization process of developed countries for hundreds of years in decades since the founding of the People’s Republic of China [1]. The technological innovation capacity of the industrial communications industry has been greatly improved, and the basic support force has been continuously strengthened, which has lifted the backbone of China’s economy. However, with the transformation and upgrading of the industrial structure, a large number of resources have flowed to the tertiary industry, and capital issues have become one of the major adverse factors restricting the development of industrial enterprises. According to the 2017 national enterprise burden survey and evaluation report, 69% of enterprises face development bottlenecks with high financing costs, which seriously restricts their operation and sustainable development. In China, many enterprises are closed every year, and of the many causes of business failures, 62% are caused by unresolved financing problems [2]. Reducing financing costs and expanding financing channels has become one of the main policy demands of most enterprises [3]. In recent years, the external business environment of China’s industrial enterprises is deteriorating, and many industrial enterprises are unable to break through the bottleneck of their development; the organizational structure is complex, and the management is not standardized. In addition, the transformation of economic structure has gradually reduced the profitability of industrial enterprises, which has brought certain difficulties to the financing of industrial enterprises. Financing constraints still affect China’s economic growth and efficiency improvement from various aspects, resulting in resource misallocation and large output and efficiency losses by inhibiting the R&D innovation of enterprises [4,5]. The differentiated financing constraint level among enterprises expands the dispersion of industrial productivity distribution, which is an important reason for resource misplacement [6]. How to break through the development bottleneck of industrial enterprises, find more financing channels and improve the financing efficiency of industrial enterprises is an urgent problem to be solved in the development of industrial enterprises at this stage. Although the government has issued a series of related policies to help enterprises alleviate financing difficulties, financing issues are still the bottlenecks that hinder the survival and development of industrial enterprises, especially for non-state-owned enterprises (NSOEs) and small and medium-sized enterprises (SMEs). They further affect China’s economic growth and efficiency improvement from various aspects, resulting in resource misallocation and large output and efficiency losses by inhibiting R&D innovation of enterprises.

Funds are the “blood” of an enterprise and the source of its vitality. Enterprise financing is the basic condition for determining the operation status of an enterprise and expanding its size. Efficiency has always been the core issue of economics research, and the low financing efficiency has always been a problem in China. Financing efficiency is reflected in two aspects. First, in the process of development, enterprises can raise the required funds in a timely and effective manner at a low cost. The second is whether the enterprise can make full use of the capital obtained from financing.

Based on the present status of financing for industrial enterprises in China, this paper divides the financing process into two processes: fund raising and fund using, combining the two-stage dynamic network slacks-based model (DNSBM) to measure the financing efficiency of China’s industrial listed enterprises. This paper aims to point out the problems of financing efficiency and the direction of improvement and put forward suggestions to improve financing efficiency.

The rest of this article is organized as follows. Section 2 is the literature review. Section 3 is the methodology and materials. Section 4 is the results and empirical analysis. Section 5 is the discussion, including policy implications, and Section 6 is the conclusions and limitations, and future research prospects.

2. Literature Review

2.1. Geographical Coverage

There are not many studies on financing efficiency in early western literature, mainly because western countries had mature property rights systems, and enterprises had high financing efficiency. Based on theoretical research, some foreign scholars have recently investigated financing efficiency from an empirical perspective. Williamson(1988) pointed out that issuing bonds and issuing stocks were the two main financing means for enterprises [7]. Meanwhile, he also analyzed the efficiency of equity and debt financing methods from the perspective of asset specificity. Some scholars specifically studied equity financing efficiency. Kunt and Levine (1996) studied research on the efficiency of the stock market and indicated that the amount of stock financing cannot fully reflect the role of the stock market. The more important role of the stock market is to improve the efficiency of capital allocation [8]. Parhankangas and Smith (2000) believed that the financing efficiency of American enterprises was affected by the financing layout of enterprises. The paper researched the financing situation of many American listed enterprises and found that when the ratio of equity financing to total funds was high, the overall financing efficiency of the enterprise would decline, so a good financing structure would improve the financing of the enterprise to a certain extent [9]. Smolarski and Kut (2011) assumed that enterprise financing efficiency is the ability to obtain maximum capital based on its appropriate allocation to achieve high earnings [10]. Balakrishnan et al. (2014) studied American companies and concluded that equity financing amount would affect the financing efficiency [11]. As for how to improve financing efficiency, Stulz (1990) argued that reducing the agency costs of managerial discretion could help to improve the efficiency of corporate finance [12]. Jayraman (2012) believed that to improve financing efficiency, enterprises could meet bank loan requirements by increasing credit and productivity [13]. However, Forrester and Reames (2020) argued even if enterprises apply for government funding or bank loans, their operating costs remain high, and their financing efficiencies are not significantly improved [14]. Gomariz and Ballesta (2014) took Spanish listed companies from 1998 to 2008 as samples. Their research showed that increasing financial leverage can reduce overinvestment, and improving the debt maturity structure would help improve financing efficiency, thus reducing overinvestment and underinvestment [15]. Cardone et al. (2005) argued that to improve their financing, SMEs could maintain longer relationships with financial entities to enjoy better access to credit, and the maintenance of banking relationships through the rendering of services reduces bank requirements in terms of guarantees in credit applications [16]. Gheeraert and Weill (2015) argued that the bank and bond financing models they employed often triggered interest rate volatility and default regulation [17].

2.2. Chronological Coverage (Last 3 Decades in China)

The term “financing efficiency” was first proposed by Chinese scholar Zeng (1993). He pointed out seven factors affecting financing efficiency and cost, but he did not define financing efficiency, which was later defined by other scholars [18]. Lu (2001) defined financing efficiency as the efficiency with which the accumulated savings are used to generate returns on investment projects [19]. Ma and Song (2004) believed that financing efficiency included two aspects: on the one hand, whether the funds raised by enterprises were used effectively; and on the other hand, whether enterprises could raise the funds they need at the lowest possible cost with the continuous improvement of the capital market [20]. Zhang and Zhao (2015) believe that financing efficiency refers to the benefits obtained from the cost paid by an enterprise in the financing process [21]. According to Wu and Huang (2021), financing efficiency is the ability of an enterprise to raise funds at the lowest cost and risk, and its ability to use those funds to maximize returns [22].In summary, the current definition of financing efficiency is mainly in the following two aspects: First, in terms of the ability to raise funds, financing efficiency is defined as the cost and risk borne by the company to raise funds. Secondly, from the perspective of fund using, financing efficiency is defined as the use efficiency of funds raised by enterprises, that is, the effective output is measured under a fixed input.

With the deepening of scholars’ research on financing efficiency, the research methods have been gradually broadened, and the initial qualitative analysis has turned to quantitative analysis, mainly using the fuzzy comprehensive evaluation method (FCE), factor analysis method, data envelopment analysis (DEA), entropy method, stochastic frontier analysis (SFA), and other research methods. At present, the most commonly used method for measuring efficiency is DEA, proposed by Charnes et al. (1978) [23], which does not need to build specific function models and has more objectivity and authenticity than expert weighting. For example, Liu et al. (2004) and Wang et al. (2006) have used DEA to measure financing efficiency [24,25]. By using the DEA model, Liu (2017) found that the equity financing efficiency of China’s growth enterprise market (GEM) companies was low, and most companies have problems with input redundancy and output shortages [26]. Wang et al. (2017) showed that the overall financing efficiency of China’s SMEs was low, only 0.445, nearly half of the enterprises’ financing efficiency lower than the average; 95% of enterprises had a financing efficiency lower than 0.643 [27]. Zeng and Geng (2019) used Super-SBM to calculate the financing efficiency difference of high-end equipment manufacturing enterprises and used the Malmquist model to study the dynamic changes of the comprehensive efficiency index trajectory, technical efficiency trajectory, and technological progress efficiency trajectory [28]. Liu et al. (2019) evaluated the financing efficiency of 85 listed companies in China’s low-carbon industries by applying a three-stage DEA [29]. Jin et al. (2021) measured the financing efficiency of listed energy conservation and environmental protection firms by using the DEA-BCC model [30].

DEA has become the mainstream trend in research efficiency. But most research uses the traditional DEA methods, such as classical CCR and BCC models, to obtain the production efficiency value. The biggest shortcoming of these methods is that they do not evaluate the impact of changes in the relevant indicator “relaxation”. In a few studies, two-stage DEA, super-DEA, three-stage DEA, and DEA-Malmquist are used, but the dynamic network DEA model is rarely found to measure financing efficiency. Moreover, in the selection of indicators, it is more scientific to consider related evaluation indicators from the perspective of capital structure in the calculation of financing efficiency. Many of the indicators selected by researchers cannot appropriately reflect the financing process and its characteristics. The selection of indicators is not highly correlated with financing, and some of the indicators even overlap with one another. Finally, many pieces of research only study financing efficiency from the static perspective, while neglecting to explore the financing contradiction caused by the efficiency change in industrial enterprises from the perspective of dynamic development.

This paper focuses on the financing efficiency of industrial listed enterprises that affects the “lifeline” of corporate survival. In terms of research methods, the two-stage DNSBM model is used to measure the financing efficiency of industrial listed enterprises. Its advantages can overcome the inaccuracy of measurement values of the traditional DEA model and the deficiency of static research.

3. Methodology and Materials

3.1. Methodology

3.1.1. Network SBM DEA

The traditional DEA model does not consider the intermediate output of the research process, but the decision-making unit (DMU) is regarded as a “black box”, focusing only on the relative efficiency between the initial input and the final output, not on its internal structure. It is impossible to derive the efficiency of the intermediate stage of the production process and the impact of each sub-stage on the overall efficiency, that is, people do not know how the input is transformed into the output, whether the operation process itself affects the overall efficiency [31].To solve this problem, Fare and Grosskopf (1996, 2007) proposed the concept of network DEA and established a specific network DEA model to decompose the production process of the enterprise into multiple stages connected by intermediate products [32,33]. The intermediate products are not only the outputs of the previous stage, but also the inputs of the latter stage; therefore, we can examine the impact of each business link on the overall efficiency of the production system. Kao and Hwang (2008) established a two-stage DEA model that is more realistically based and fully considers the relevance of each sub-stage. The current network DEA model mainly includes chain and parallel network structures [34]. This paper adopts the two-stage network DEA model with a chain structure.

Tone and Tsutsui developed a network DEA model using a slacks-based measure called “NSBM”, it effectively overcomes the limitation of the original network DEA model to assume that all input and output factors change in the same proportion. In the NSBM model, it is assumed that there are $n$ $DMUs\left(j=1,2,\dots ,n\right)$ composed of $k\left(k=1,2,\dots ,K\right)$ dimensions. Let $m_k$ and $r_k$ be the number of inputs and outputs of dimension $k$, respectively. Tone and Tsutsui (2009) defined the efficiency of each divisional efficiency as follows [35]:

$${\rho }_k=\frac{1-\frac{1}{m_k}\left({{\displaystyle \sum }}_{i=1}^{m_k}\frac{s_i^{k-*}}{x_{io}^k}\right)}{1+\frac{1}{r_k}\left({{\displaystyle \sum }}_{r=1}^{r_k}\frac{s_r^{k+*}}{y_{ro}^k}\right)}\left(k=1,\dots ,K\right)$$

(1)

Overall efficiency is:

$${\rho }_o^{*}=\underset{{\lambda }^k,s^{k-},s^{k+}}{\min }\frac{{{\displaystyle \sum }}_{k=1}^k{w}^k\left[1-\frac{1}{m_k}\left({{\displaystyle \sum }}_{i=1}^{m_k}\frac{s_i^{k-}}{x_{io}^k}\right)\right]}{{{\displaystyle \sum }}_{k=1}^k{w}^k\left[1+\frac{1}{r_k}\left({{\displaystyle \sum }}_{r=1}^{r_k}\frac{s_r^{k+}}{y_{ro}^k}\right)\right]}$$

(2)

subject to

$${{\displaystyle \sum }}_{k=1}^k{w}^k=1,$$

$$x_0^k=X^k{\lambda }^k+s^{k-}\left(k=1,\cdots ,K\right)$$

$$y_0^k=Y^k{\lambda }^k-s^{k+}\left(k=1,\cdots ,K\right),$$

$$\mathrm{e}{\lambda }^k=1\left(k=1,\cdots ,K\right)$$

$$w^k\ge 0, {\lambda }^k\ge 0,s^{k-}\ge 0, s^{k+}\ge 0, \left(\forall k\right)$$

$$Z^{\left(k,h\right)}{\lambda }^h=Z^{\left(k,h\right)}{\lambda }^k,\left(\forall \left(k,h\right)\right)$$

where $w$ is the division weight, and $s^{k+}$ are slacks of the input and output, and $Z^{\left(k,h\right)}$ is the intermediate product between division $k$ and division $h$.

3.1.2. Dynamic DEA Model

The measurement of intertemporal efficiency changes has been a concern of the DEA. For the measurement of intertemporal efficiency changes, Färe and Grosskopf’s (1997) dynamic DEA model incorporates carry-over activities into the measurement of dynamic efficiency for the first time [36]. Bogetoft et al. (2008), Kao (2008), and Sueyoshi and Sekitani (2005) have further developed the dynamic DEA [37,38,39]. This connection relation of carry-over variables is similar to the intermediate variable in the network DEA model. The difference is that the carry-over variables are associated with two periods, and the intermediate variables are associated with two divisions. Tone and Tsutsui (2010) further introduced the SBM model into the dynamic DEA model and divided the intertemporal activity variables into four categories: good, bad, discretionary, and non-discretionary, and established non-radial and non-oriented dynamic SBM models (DSBM) [40].

3.1.3. Dynamic Network SBM (DNSBM)

To better measure the dynamic changes in department efficiency, Tone and Tsutsui (2014) further combined NSBM and DSBM, taking into account the connectivity variables and intertemporal activities, and established the dynamic network SBM(DNSBM), in which the divisions are connected by links, and two succeeding periods are combined by carry-overs. The combined model not only provides the overall efficiency during the entire observation period but can further analyze the dynamic changes of the periodic efficiency and the divisional efficiency [41].

Suppose there are $n$ $DMUs\left(j=1,2,\dots ,n\right)$ composed of $k\left(k=1,2,\dots ,K\right)$ dimensions over $T\left(t=1,\dots ,T\right)$ periods. Let $m_k$ be the inputs of division $k$, and $r_k$ be outputs to division $k$. Denote the link ($k,h$) leading from division $k$ to division $h$, and the set of links by $L^{\left(k,h\right)}$. The objective functions are as follows [41].

The overall efficiency is evaluated by the following program:

$${\theta }_0^{*}=\min \frac{{{\displaystyle \sum }}_{t=1}^T{W}^t\left[{{\displaystyle \sum }}_{k=1}^K{w}^k\left[1-\frac{1}{m_k+linki{n}_k+nba{d}_k}\left({{\displaystyle \sum }}_{i=1}^{m_k}\frac{s_{iok}^{t-}}{x_{iok}^t}+{{\displaystyle \sum }}_{{(kh)}_l=1}^{linki{n}_k}\frac{s_{o{\left(kh\right)}_{l}in}^t}{z_{o{\left(kh\right)}_{l}in}^t}+{{\displaystyle \sum }}_{k_l=1}^{nba{d}_k}\frac{s_{o{k}_{l}bad}^{\left(t,t+1\right)}}{z_{o{k}_{l}bad}^{\left(t,t+1\right)}}\right)\right]\right]}{{{\displaystyle \sum }}_{t=1}^T{W}^t\left[{{\displaystyle \sum }}_{k=1}^K{w}^k\left[1+\frac{1}{r_k+linkou{t}_k+ngoo{d}_k}\left({{\displaystyle \sum }}_{r=1}^{r_k}\frac{s_{rok}^{t+}}{y_{rok}^t}+{{\displaystyle \sum }}_{{(kh)}_l=1}^{linkou{t}_k}\frac{s_{o{\left(kh\right)}_{l}out}^t}{z_{o{\left(kh\right)}_{l}out}^t}+{{\displaystyle \sum }}_{k_l=1}^{ngoo{d}_k}\frac{s_{o{k}_{l}good}^{\left(t,t+1\right)}}{z_{o{k}_{l}good}^{\left(t,t+1\right)}}\right)\right]\right]}$$

(3)

Period efficiency is defined by:

$${\tau }_o^{t*}=\frac{{{\displaystyle \sum }}_{k=1}^K{w}^k\left[1-\frac{1}{m_k+linki{n}_k+nba{d}_k}\left({{\displaystyle \sum }}_{i=1}^{m_k}\frac{s_{iok}^{t-}}{x_{iok}^t}+{{\displaystyle \sum }}_{{(kh)}_l=1}^{linki{n}_k}\frac{s_{o{\left(kh\right)}_{l}in}^t}{z_{o{\left(kh\right)}_{l}in}^t}+{{\displaystyle \sum }}_{k_l=1}^{nba{d}_k}\frac{s_{o{k}_{l}bad}^{\left(t,t+1\right)}}{z_{o{k}_{l}bad}^{\left(t,t+1\right)}}\right)\right]}{{{\displaystyle \sum }}_{k=1}^K{w}^k\left[1+\frac{1}{r_k+link{\mathit{out}}_k+ngoo{d}_k}\left({{\displaystyle \sum }}_{r=1}^{r_k}\frac{s_{rok}^{t+}}{y_{rok}^t}+{{\displaystyle \sum }}_{{(kh)}_l=1}^{linkou{t}_k}\frac{s_{o{\left(kh\right)}_{l}out}^t}{z_{o{\left(kh\right)}_{l}out}^t}+{{\displaystyle \sum }}_{k_l=1}^{ngoo{d}_k}\frac{s_{o{k}_{l}good}^{\left(t,t+1\right)}}{z_{o{k}_{l}good}^{\left(t,t+1\right)}}\right)\right]}\left(\forall t\right)$$

(4)

where variables on the right-hand side indicate optimal values for the overall efficiency ${\theta }_0^{*}$. Divisional efficiency is defined by:

$${\delta }_{ok}^{*}=\frac{{{\displaystyle \sum }}_{k=1}^T{w}^t\left[1-\frac{1}{m_k+linki{n}_k+nba{d}_k}\left({{\displaystyle \sum }}_{i=1}^{m_k}\frac{s_{iok}^{t-}}{x_{iok}^t}+{{\displaystyle \sum }}_{{(kh)}_l=1}^{linki{n}_k}\frac{s_{o{\left(kh\right)}_{l}in}^t}{z_{o{\left(kh\right)}_{l}in}^t}+{{\displaystyle \sum }}_{k_l=1}^{nba{d}_k}\frac{s_{o{k}_{l}bad}^{\left(t,t+1\right)}}{z_{o{k}_{l}bad}^{\left(t,t+1\right)}}\right)\right]}{{{\displaystyle \sum }}_{k=1}^T{w}^t\left[1+\frac{1}{r_k+link{\mathit{out}}_k+ngoo{d}_k}\left({{\displaystyle \sum }}_{r=1}^{r_k}\frac{s_{rok}^{t+}}{y_{rok}^t}+{{\displaystyle \sum }}_{{(kh)}_l=1}^{linkou{t}_k}\frac{s_{o{\left(kh\right)}_{l}out}^t}{z_{o{\left(kh\right)}_{l}out}^t}+{{\displaystyle \sum }}_{k_l=1}^{ngoo{d}_k}\frac{s_{o{k}_{l}good}^{\left(t,t+1\right)}}{z_{o{k}_{l}good}^{\left(t,t+1\right)}}\right)\right]}\left(\forall t\right)$$

(5)

Finally, period-divisional efficiency is defined by:

$${\rho }_{ok}^{t*}=\frac{1-\frac{1}{m_k+linki{n}_k+nba{d}_k}\left({{\displaystyle \sum }}_{i=1}^{m_k}\frac{s_{iok}^{t-}}{x_{iok}^t}+{{\displaystyle \sum }}_{{(kh)}_l=1}^{linki{n}_k}\frac{s_{o{\left(kh\right)}_{l}in}^t}{z_{o{\left(kh\right)}_{l}in}^t}+{{\displaystyle \sum }}_{k_l=1}^{nba{d}_k}\frac{s_{o{k}_{l}bad}^{\left(t,t+1\right)}}{z_{o{k}_{l}bad}^{\left(t,t+1\right)}}\right)}{1+\frac{1}{r_k+link{\mathit{out}}_k+ngoo{d}_k}\left({{\displaystyle \sum }}_{r=1}^{r_k}\frac{s_{rok}^{t+}}{y_{rok}^t}+{{\displaystyle \sum }}_{{(kh)}_l=1}^{linkou{t}_k}\frac{s_{o{\left(kh\right)}_{l}out}^t}{z_{o{\left(kh\right)}_{l}out}^t}+{{\displaystyle \sum }}_{k_l=1}^{ngoo{d}_k}\frac{s_{o{k}_{l}good}^{\left(t,t+1\right)}}{z_{o{k}_{l}good}^{\left(t,t+1\right)}}\right)}\left(\forall k;\forall t\right)$$

(6)

If there is no desirable output and undesirable output, the program contains no corresponding variables for “bad” and “good”. The variables in the formulas are shown in

Table 1. Data and variables.

Data	Variable	Definition
Input	$x_{ijk}^t$	Input resource $i$ to $DM{U}_j$ for division k at period $t.$
Output	$y_{rjk}^t$	Output product $r$ from $DM{U}_j$ for division $k$ at period $t$.
Link	$z_{ij\left(kh\right)l}^t$	$DM{U}_j$ from division $k$ to division $h$ at period $t$.
Carry-over	$z_{jkl}^{\left(t,t+1\right)}$	Carry-over of $DM{U}_j$ at division $k$ from period $t$ to period $t+1$.
Input slack	$s_{iok}^{t-1}$	Slack of input $i$ of $DM{U}_o$ for division $k$ at period $t$.
Output slack	$s_{rok}^{t+}$	Slack of output r of $DM{U}_o$ for division $k$ at period $t$.
Link slack	$s_{o{\left(kh\right)}_{l}a}^t$	Slack of link ${\left(kh\right)}_l$ of $DM{U}_o$ at period $t$.a stands for free, as-input, and as-output.
Carry-over slack	$s_{o{k}_{l}a}^{\left(t,t+1\right)}$	Slack of carry-over $k_l$ from period $t$ o period $t+1$.
Intensity	${\lambda }_{jk}^t$	Intensity of $DM{U}_j$ corresponding to division $k$ at period $t$.

.

3.2. Materials

3.2.1. Data Sources

Since China’s reform and opening-up, China’s industry sector has been developing rapidly. According to the data of the World Bank, measured in current dollars, China’s manufacturing value added exceeded that of the United States for the first time in 2010 and became the largest manufacturing country in the world. Since then, China’s manufacturing value added has accounted for 27.0% of the world’s share in 2017, becoming an important engine driving global industrial growth [42]. China’s industry maintained a good growth rate from 2010 to 2017. Based on this background, this paper takes China’s A-share industrial listed enterprises as samples with a 7-year research window from 2011 to 2017. To ensure the acquisition of variable lags terms, the actual selection of Chinese industrial listed enterprises is from 2010 to 2017. To maintain the consistency of the data, the sample selection used the following criteria: (1) Exclude ST (special treatment) enterprises. ST enterprises refer to listed enterprises that have been specially treated by the stock exchange due to financial abnormalities or other reasons; (2) exclude enterprises with zero equity cost; (3) exclude enterprises that issued an IPO in that year; (4) exclude listed enterprises that lack relevant data; (5) delete listed enterprises in Tibet. The final sample consists of 450 enterprises in 29 provinces from 2010 to 2017. The data comes from the China Stock Market & Accounting Research (CSMAR) Database.

3.2.2. DEA Indicator Selection

This paper divides financing efficiency into two stages. The first stage is fund raising, which includes external financing and internal financing. External financing mainly includes debt financing and equity financing. The financing cost is taken as an input and the obtained financing amount is taken as an output, which reflects the fund raising efficiency; the second stage is fund using. The amount of external financing obtained and the retained earnings from the previous period are used in the current period. The performance of the enterprise is used as an output to reflect the allocation efficiency of corporate financing. The selection of input-output indicators in this paper mainly refers to Tan et al. (2019) [43].

• Input indicators:

Debt financing cost is the cost paid by companies to raise debts and is expressed by the financial expense indicator in financial statements.

Equity financing cost is the capital asset pricing model (CAPM) and is used to measure the equity financing cost. The CAPM model is: $R_i=R_f+{\beta }_i\left(R_m-R_f\right)$, where $R_i$ reflects the equity cost, and the rate of return of stock $i$; $R_f$ is risk-free rate; $R_m$ is expected market return; and ${\beta }_i$ is systemic risk of asset $i$, so, is the equity market premium.

2.

Intermediate indicators:

Total debt financing refers to the current obligation formed by past transactions and events, including loans from financial institutions and the issuance of bonds, etc.

Total equity financing is the net cash flow received from the issuance of shares, including equity and equity premium. Share premium is the amount of money actually received by a corporation for issuing shares at a premium over the total book value of the shares.

3.

Carry-over:

Internal financing includes retained earnings, undistributed profits, and accumulated depreciation. Its advantages are a low cost, almost zero; it can independently determine the use of funds; and it has strong autonomy.

4.

Output indicators:

Economic value added (EVA) refers to the balance of net profit after-tax, and after deducting all the capital costs invested by an enterprise. It is a business performance assessment tool used to reflect the ultimate business objectives of the enterprise.

Tobin’s Q is the ratio of a company’s market value to its asset replacement cost. If the Q value is high, the company will issue fewer stocks and buy more investment products, so that the investment cost will increase. When the reverse is done the investment cost will decrease.

Weighted return on equity (ROE) is an important indicator to judge the profitability of listed companies. It can measure how efficiently a company is using capital invested by shareholders.

Main business revenue growth rate (MBRG) refers to the ratio of the increase in the current year’s operating revenue to the total operating revenue of the previous year. It indicates the increase and decrease in main business income compared with the previous year.

Table 2. Indicators of the two-stage DNSBM DEA model of financing efficiency.

Indicator		Indicator Definition	Units
Input	Debt financing cost	The cost of fund raising	yuan
	Equity financing cost	$R_i=R_f+{\beta }_i\left(R_m-R_f\right)$	yuan
Intermediate	Total debt financing	Short-term loans+ bonds payable +long-term liabilities maturing in one-year +long-term loans	yuan
	Total equity financing	Equity + equity premium	yuan
Carry-over	Internal financing	Retained earnings + undistributed profits +accumulated depreciation	yuan
Output	Economic value added (EVA)	Net operating profit after tax (NOPAT)–capital cost	yuan
	Tobin’s Q	Market value/asset replacement cost	-
	Return on equity (ROE)	Net profit/average net assets	-
	Main business revenue growth rate (MBRG)	(Revenue growth/total revenue of last year) × 100%	-

shows the indicators and their definitions of the two-stage DNSBM model of financing efficiency.

According to the indicators selected above, this paper constructs the two-stage DNSBM-DEA model shown in Figure 1. In Figure 1, the financing of each period has two stages: the fund raising stage and the fund using stage.

In the fund raising stage (the first stage) of period t, debt financing cost and equity financing cost are the sources of funds, used as input indicators. Total debt financing and total equity financing, as intermediate products, are not only the outputs of the fund raising stage, but also the inputs of the fund using stage. EVA, Tobin’s Q, ROE, and MBRG reflect the performance of the enterprise and are the outputs of the fund using stage (the second stage).

In period t + 1, the input indicators and output indicators of the first stage are the same as in period t, period t’s retained earnings, etc., can be seen as internal financing, which increases the amount of funds used in period t + 1. Therefore, the input indicator of the fund using stage (the second stage) of period t + 1 increased internal financing, which, as a carry-over, connects period t and t + 1.

4. Results

This paper utilizes the two-stage DNSBM model of MaxDEA 8 Ultra software to calculate the comprehensive technical efficiency of the sample company. In theory, the DEA model requires that inputs and outputs should not be negative, otherwise the result is invalid. Therefore, the original data are normalized, which can not only avoid the negative indicators from affecting the accuracy of the model evaluation but also fundamentally eliminate the defects of the indicators based on maintaining the model processing results. This paper draws on the empirical methods of Xiong et al. (2014) and Fang et al. (2015) [44,45] and uses the following formula to process the data:

$$X_{\mathrm{norm}}=0.1+\frac{X-X_{min}}{X_{max}-X_{min}}\times 0.9$$

(7)

where $X_{\mathrm{norm}}$ is the normalized data,$X$ is the original data, and $X_{max}$ and $X_{min}$ are the maximum and minimum values of the original data set, respectively.

It can be seen from

Table 3. Results of the two-stage DNSBM model.

Year	Score	Mean	Max	Min	Score = 1	0.5 ≤ Score < 1	0 ≤ Score < 0.5
2011–2017	θ	0.598	1	0.203	4 (0.89%)	375 (83.33%)	71 (15.78%)
2011	θ1	0.612	1	0.199	19 (4.22%)	396 (88.00%)	35 (7.78%)
	θ2	0.604	1	0.022	113 (25.11%)	164 (36.44%)	173 (38.44%)
	θ^	0.608	1	0.161	19 (4.22%)	287 (63.78%)	144 (32.00%)
2012	θ1	0.551	1	0.080	18 (4.00%)	290 (64.44%)	142 (31.56%)
	θ2	0.622	1	0.026	88 (19.56%)	228 (50.67%)	134 (29.78%)
	θ^	0.586	1	0.079	18 (4.00%)	287 (63.78%)	145 (32.22%)
2013	θ1	0.614	1	0.159	26 (5.78%)	359 (79.78%)	65 (14.44%)
	θ2	0.756	1	0.063	126 (28.00%)	250 (55.56%)	74 (16.44%)
	θ^	0.685	1	0.156	26 (5.78%)	351 (78.00%)	73 (16.22%)
2014	θ1	0.663	1	0.206	21 (4.67%)	389 (86.44%)	40 (8.89%)
	θ2	0.685	1	0.037	81 (18.00%)	281 (62.44%)	88 (19.56%)
	θ^	0.674	1	0.213	21 (4.67%)	349 (77.56%)	80 (17.78%)
2015	θ1	0.615	1	0.224	22 (4.89%)	380 (84.44%)	48 (10.67%)
	θ2	0.435	1	0.019	53 (11.78%)	121 (26.89%)	276 (61.33%)
	θ^	0.525	1	0.167	22 (4.89%)	177 (39.33%)	251 (55.78%)
2016	θ1	0.574	1	0.272	17 (3.78%)	262 (58.22%)	171 (38.00%)
	θ2	0.558	1	0.019	82 (18.22%)	164 (36.44%)	204 (45.33%)
	θ^	0.566	1	0.179	17 (3.78%)	231 (51.33%)	202 (44.89%)
2017	θ1	0.558	1	0.138	14 (3.11%)	298 (66.22%)	138 (30.67%)
	θ2	0.512	1	0.017	67 (14.89%)	155 (34.44%)	228 (50.67%)
	θ^	0.535	1	0.137	14 (3.11%)	228 (50.67%)	208 (46.22%)

Note: θ is the overall financing efficiency for seven years, θ¹ is the fund raising efficiency in the first stage, θ² is the fund using efficiency score in the second stage, and θ^{^} is the annual overall financing efficiency. The number of enterprises is presented outside the brackets and the percentage is inside the brackets.

and Figure 2 that the overall financing efficiency θ of Chinese industrial enterprises in the seven years from 2011 to 2017 is low, with an average value of 0.598, a maximum value of 1, and a minimum value of 0.203. Among the 450 samples, only 4 enterprises have an overall financing efficiency value of 1, which has reached technical efficiency, accounting for less than 1%, indicating that the fund raising efficiency and fund using efficiency of these 4 enterprises reached technical efficiency from 2011 to 2017. In the overall sample, there were 375 enterprises with an efficiency value between 0.5 and 1, accounting for 83.33%, and 71 enterprises with an efficiency value less than 0.5, accounting for 15.78%.

Considering the two-stage financing efficiency from 2011 to 2017, the fund raising efficiency of the first stage reached the optimal level in 2014. Before 2014, the fund raising efficiency is lower than the fund using efficiency of the second stage, and after 2014 it is reversed. After falling in 2012, fund raising efficiency continued to rise in 2013 and 2014 but has continued to decline after 2014. One of the reasons is that after the bear market in 2011 and 2012, China’s stock market gradually rose in volatility from 2013 to the first half of 2014 and began to enter a new bull market in the second half of 2014, until the stock market collapsed in June 2015. These stock market fluctuations had a great influence on the enterprise stock financing cost. On the contrary, the basic trend of the fund using efficiency is similar to that of overall financing efficiency per year. After reaching its highest point in 2013, it shows a downward trend and reaches its lowest point in 2015. The fluctuation range of the fund using efficiency of the second stage is generally greater than that of the fund raising efficiency. In 2013, the profits of industrial enterprises above the designated size in China showed a rapid growth trend, and the profits of most industries increased over the previous year [46]. In 2014, although the operating income and profits of Chinese industrial-listed enterprises maintained growth, the growth rate declined [47]. In 2015, insufficient demand led to a significant slowdown in production and sales [48]. The price of industrial products decreased significantly. High costs and tight liquidity restrict the production and operation of enterprises, which are the main reasons for the decline in profits of industrial enterprises [49]. From 2016 to 2017, with the deepening of supply-side structural reform, positive results were achieved in cutting capacity, deleveraging, and reducing cost in the industrial sector, with significant improvement in operating efficiency and a further improvement in operation quality [50].

Table 4. Overall financing efficiency of different types of enterprises.

Statistics Type		Mean	Max	Min	Score = 1	0.5 ≤ Score < 1	0 ≤ Score < 0.5
All samples (450)		0.598	1	0.203	4 (0.89%)	375 (83.33%)	71 (15.78%)
Property	SOEs (232)	0.589	1	0.203	4 (1.72%)	179 (77.16%)	49 (21.12%)
	NSOEs (218)	0.608	0.940	0.322	0 (0%)	196 (89.91%)	22 (10.09%)
Listed board	Main board (294)	0.581	1	0.203	4 (1.36%)	225 (76.53%)	65 (22.11%)
	SME board (134)	0.623	0.893	0.416	0 (0.00%)	128 (95.52%)	6 (4.48%)
	GEM board (22)	0.674	0.929	0.539	0 (0.00%)	22 (100.00%)	0 (0.00%)
Region	Eastern (292)	0.606	1	0.203	3 (1.03%)	248 (84.93%)	41 (14.04%)
	Central (96)	0.590	1	0.361	1 (1.04%)	77 (80.21%)	18 (18.75%)
	Western (62)	0.573	0.821	0.324	0 (0.00%)	50 (80.65%)	12 (19.35%)

Note: On 6 April 2021, the main board of the Shenzhen Stock Exchange and the SME board were formally merged. After the merger, the conditions for IPO, investor threshold trading mechanism, and the securities code and abbreviation remain unchanged. Investors can identify SMEs according to the securities code.

shows the overall financing efficiency of different types of enterprises. We can see that the four enterprises that are effective in overall financing efficiency technology are state-owned enterprises (SOEs) listed on the main-board market, of which three are in the eastern region and one is in the central region. There are obvious differences in the financing efficiency values of different types of enterprises.

Table 4 and

Table 5. Results of two-stage financing efficiency of different property listed companies from 2011 to 2017.

	Score	SOEs			NSOEs
Year		θ1	θ2	θ^	θ1	θ2	θ^
2011		0.605	0.543	0.574	0.620	0.668	0.644
2012		0.526	0.589	0.553	0.578	0.656	0.622
2013		0.619	0.729	0.653	0.609	0.785	0.719
2014		0.655	0.679	0.655	0.673	0.690	0.694
2015		0.614	0.418	0.516	0.617	0.452	0.535
2016		0.571	0.520	0.543	0.577	0.599	0.590
2017		0.552	0.488	0.519	0.564	0.537	0.552
Mean		0.592	0.567	0.573	0.605	0.627	0.622

show that the overall financing efficiency of SOEs is lower than that of NSOEs, mainly because more enterprises have an overall financing efficiency of less than 0.5. The average two-stage financing efficiency of SOEs from 2011 to 2017 and the average overall financing efficiency of each year are also lower than those of NSOEs. Among them, SOEs’ average annual fund using efficiency is 0.567, while the NSOEs’ average annual fund using efficiency is 0.627, and the gap (0.06) is significantly larger than the gap (0.013) in the first stage of fund raising efficiency.

The average fund raising efficiency of SOEs is greater than the fund using efficiency, but the opposite is true for NSOEs. The possible reason is that, on the one hand, due to the additional social burdens of economic growth, stable employment, and protection of financial income, local governments usually grant certain “implicit guarantees” as “compensatory considerations” (Rao and Jiang, 2013; Wang et al.,2015) [51,52]. Therefore, it is easier for SOEs to obtain bank loans. Since China’s large-scale commercial banks are state-owned holdings, they are in the same category as SOEs in terms of property rights, which naturally causes subjective and emotional “credit discrimination” to private enterprises (Lu et al., 2009) [53]. On the other hand, due to the particularity of the property rights and economic status of SOEs, inappropriate government intervention will not only reduce the information asymmetry and transaction cost of banks and other institutions to ease financing constraints, but also increase the information asymmetry between controlling shareholders and executives, and further increase the transaction cost and agency cost, thereby reducing the fund using efficiency (Lai et al., 2019) [54]. Therefore, there is a deviation between the financing advantages of SOEs and the fund using efficiency. This indicates that whether SOEs have financing advantages or not, their fund using inefficiency is a sufficient fact. Although the fund using efficiency of NSOEs is higher than their fund raising efficiency, the investment activities of fund using are still greatly restricted by the financing constraints, so the overall financing efficiency is not high.

Table 4 and

Table 6. Results of the two-stage financing efficiency of listed companies in different listed boards from 2011 to 2017.

	Score	Main Board			SME Board			GEM Board
Year		θ1	θ2	θ^	θ1	θ2	θ ^	θ1	θ2	θ^
2011		0.596	0.521	0.559	0.639	0.733	0.686	0.661	0.924	0.792
2012		0.514	0.558	0.536	0.613	0.715	0.664	0.671	0.906	0.788
2013		0.607	0.747	0.652	0.613	0.763	0.735	0.717	0.837	0.825
2014		0.653	0.674	0.652	0.677	0.700	0.712	0.725	0.732	0.740
2015		0.608	0.417	0.512	0.625	0.452	0.539	0.653	0.569	0.611
2016		0.575	0.511	0.537	0.574	0.625	0.606	0.572	0.776	0.713
2017		0.548	0.489	0.514	0.566	0.539	0.561	0.643	0.656	0.649
Mean		0.586	0.559	0.566	0.615	0.647	0.643	0.663	0.771	0.731

show that the overall financing efficiency of listed enterprises on the main board is the lowest, which may be mainly due to the large proportion of SOEs, 205 of 294 enterprises are SOEs, and that of the listed enterprises on the GEM board is the highest. The overall financing efficiency of all listed enterprises on the GEM board is greater than 0.5. The average two-stage financing efficiency of GEM listed enterprises from 2011 to 2017 and the average financing efficiency of each year are significantly higher than those of the main board and SME board listed enterprises, and the financing efficiency values on the main board are the lowest. The most obvious gap is in the second stage of fund using, but this gap is gradually narrowing. The fund using efficiency of listed enterprises on the main board and SME board reached the highest in 2013, while the financing efficiency of GEM enterprises reached the highest in 2011, which may be related to the fact that GEM is favored by funds shortly after its establishment and the outstanding performance of listed enterprises. The average fund raising efficiency of the main-board-listed enterprises in the first stage (0.586) is greater than the average fund using efficiency in the second stage (0.559). On the contrary, the fund using efficiency of SMEs and GEM listed enterprises is greater than the fund raising efficiency.

China has a vast territory and a large population. There are great differences among provinces and cities in the natural conditions, historical basis, social and economic development level, and national policies resulting in significant regional development differences. Figure 3 shows some characteristics of the 29 provinces, and among the 29 provinces, except for Fujian, Guangxi, Inner Mongolia, Gansu, Ningxia, Sichuan, Xinjiang, Yunnan and Qinghai where the main land use is garden land, the main land use of other provinces is land for inhabitation, mining and manufacturing [55]. From the perspective of the region, Table 4 and

Table 7. Results of the two-stage financing efficiency of listed companies in different regions from 2011 to 2017.

	Score	Eastern Region			Central Region			Western Region
Year		θ1	θ2	θ^	θ1	θ2	θ^	θ1	θ2	θ^
2011		0.617	0.633	0.625	0.599	0.591	0.595	0.609	0.487	0.548
2012		0.565	0.650	0.605	0.520	0.603	0.569	0.532	0.519	0.527
2013		0.616	0.785	0.699	0.619	0.696	0.684	0.596	0.716	0.620
2014		0.666	0.695	0.682	0.669	0.671	0.673	0.641	0.657	0.635
2015		0.621	0.459	0.540	0.618	0.443	0.531	0.583	0.309	0.446
2016		0.585	0.580	0.583	0.554	0.537	0.549	0.556	0.488	0.514
2017		0.563	0.527	0.547	0.557	0.500	0.525	0.534	0.462	0.494
Mean		0.605	0.618	0.612	0.591	0.577	0.589	0.579	0.520	0.541

show that the overall financing efficiency of Chinese industrial enterprises has obvious regional characteristics. The overall financing efficiency of the listed enterprises in the eastern region is significantly higher than the central and western regions, and the central region is also significantly better than the western region. This is mainly reflected in the finding that the proportion of enterprises with an efficiency score below 0.5 is higher than that in the eastern region, and only one of the regional listed enterprises has an overall financing efficiency greater than 0.8. From the perspective of the two-stage financing efficiency from 2011 to 2017, the fund raising efficiency value in each region is not much different, while the fund using is significantly different, showing an overall pattern of high levels in the east and low levels in the west; in particular, the financing efficiency gap is large in the eastern region between western region.

Table 8. Average financing efficiency of industrial listed companies in different provinces.

Region	Province	Score	Region	Province	Score	Region	Province	Score
Eastern region (292)	Beijing (30)	0.610	Central region (96)	Anhui (12)	0.660	Western region (62)	Gansu (7)	0.534
	Fujian (10)	0.615		Henan (18)	0.534		Guizhou (7)	0.610
	Guangdong (63)	0.618		Heilongjiang (3)	0.551		Ningxia (2)	0.570
	Guangxi (5)	0.501		Hubei (14)	0.662		Qinghai (1)	0.502
	Hebei (13)	0.651		Hunan (13)	0.626		Shaanxi (6)	0.589
	Jiangsu (45)	0.599		Jilin (9)	0.551		Sichuan (15)	0.608
	Liaoning (9)	0.591		Jiangxi (14)	0.601		Xinjiang (11)	0.554
	Shandong (38)	0.627		Inner Mongolia (8)	0.499		Yunnan (7)	0.567
	Shanghai (25)	0.561		Shanxi (5)	0.643		Chongqing (6)	0.576
	Tianjin (9)	0.590
	Zhejiang (45)	0.612

shows the overall situation of the financing efficiency of listed enterprises in each province. All the samples in this paper cover listed enterprises in 29 provinces in China, and the average overall financing efficiency of 14 provinces is lower than the national level (0.598), accounting for 48.276%, among them, 4 out of 11 provinces are in the eastern region, 4 out of 9 provinces are in the central region, and 7 out of 9 provinces are in the western region. This is consistent with China’s regional heterogeneity. The three provinces with the lowest overall financing efficiency are Inner Mongolia, Qinghai, and Guangxi. Inner Mongolia is an autonomous region of agriculture and animal husbandry, coal is the leading industrial product, and the processing and manufacturing industry is developing slowly. There is only one sample enterprise in the Qinghai Province, and its financing efficiency is low; to a certain extent, it reflects the economic development of industrial enterprises in the Qinghai Province. Guangxi is an underdeveloped province in China. The capital shortage, which is slowing down the growth of industrial investment, along with other factors, restricts the development of Guangxi industry [56].

5. Discussion

This paper has provided an empirical assessment of the dynamic financing efficiency of 450 A-share industrial listed enterprises from 2011 to 2017 by using the two-stage dynamic network SBM(DNSBM) model. We found that: (1) The overall financing efficiency of China’s industrial enterprises from 2011 to 2017 is low, with an average score of 0.598. Only 4 of the 450 samples have an overall financing efficiency score of 1, accounting for less than 1%. From each year, both fund raising efficiency and fund using efficiency are not high, mainly between 0.5 and 0.7. Most enterprises have redundant capital investment and allocation, the funds have not been effectively used, and there is much room for improvement in the financing efficiency of Chinese industrial enterprises. (2) The overall financing efficiency, fund raising efficiency, and fund using efficiency of SOEs are all lower than 0.6 and lower than that of NSOEs. Previous studies have reached similar conclusions, for example, Tan et al. (2019) believed that the fund using efficiency of SOEs is much lower than that of private companies [43]. Jin (2015) and Li (2017) concluded that the total factor productivity (TFP) of SOEs was generally low. In addition, an interesting finding is that the average fund raising efficiency of SOEs is greater than the fund using efficiency, but the opposite is true for NSOEs. Compared with SOEs, the financing cost of NSOEs was not only higher, but also more unbalanced [5,57]. (3) The overall financing efficiency of the main-board-listed enterprises is the lowest with a score of 0.581, and that of the growth enterprise market (GEM) is the highest with a score of 0.674. This is mainly reflected in the fact that the average fund using efficiency (0.559) of the second stage of the main-board enterprises is lower than the average fund raising efficiency (0.586) of the first stage; however, for the SMEs and the GEM enterprises, the average fund using efficiency of the second stage is much higher than the average fund raising efficiency of the first stage. This is consistent with the research conclusion of Yi et al. (2021); that is, the financing efficiency of the second board (i.e., GEM) is higher than that of the other two boards (SMEs board and main board) [30]. (4) The overall financing efficiency of China’s industrial enterprises has obvious regional characteristics. All financing efficiency scores of enterprises in the eastern region are higher than 0.6, while those in the central and western regions are lower than 0.6. The average fund using efficiency (0.618) is higher than the average fund raising efficiency (0.605) in the eastern region, while the opposite is true in the central and western regions. In the central and western regions, the fund raising efficiency of industrial enterprises is not high, and the fund using efficiency is lower. This is consistent with China’s regional heterogeneity.

The improvement of financing efficiency is of great significance for promoting China’s industrial revitalization and achieving high-quality industrial development. Financing efficacy helps enterprises obtain financial support for stimulating capital accumulation (Ma et al., 2020) [58], improves distribution efficiency, and promotes the healthy and stable development of enterprises (Yi et al., 2022) [59].To improve the financing efficiency of China’s A-share industrial listed enterprises, on the one hand, China’s industrial listed enterprises must improve their financing channels, choose the best financing method, and maintain a reasonable debt financing ratio, on the other hand, it is necessary to improve the operation and management level and profitability, increase the value of the enterprise, enhance the debt-paying ability, and attract more capital at low cost. According to the characteristics of enterprises’ financing efficiency, the government can adopt differentiated assistance policies: (1) For SOEs, they should mainly improve the fund using efficiency. The government should further deepen the reform of SOEs, streamline government institutions, and delegate more management power to enterprises. This will facilitate SOEs to enter the market and participate in the competition, which will in turn change the operating mechanism, improve the efficiency of resource utilization, and further stimulate the vitality of SOEs. It may be necessary to refine corporate governance structures in both SOEs and NSOEs, involving a more active board of directors or constraints on the use of internal funds (He et al., 2019) [60]. (2) In China, most of the enterprises are small enterprises, and more than 95% of them are NSOEs. NSOEs are the biggest contributors to government tax and national financial resources, with tax contributions exceeding 50%. However, due to the impact of financing constraints, NSOEs, especially SMEs, have difficulty in financing [61]. The sample of this study is limited to industrial listed enterprises. Non-listed NSOEs and SMEs cannot raise funds in the stock market and have fewer financing channels and more financing restrictions. Therefore, NSOEs and SMEs should focus on improving the fund raising efficiency by easing financing constraints. It is necessary to improve the government financing services, construct a multi-level capital market system, encourage innovation in financial products and businesses, continuously expand financing channels, increase guarantee methods, and ensure various types of enterprises can all get timely, full, and low-cost development funds. Supply chain finance helps to effectively alleviate the financing constraints of SMEs (Cai, 2021) [62]. The government should create a good institutional environment for the development of supply chain finance, and the preferential policies of bank credit support should be more inclined to non-state-owned SMEs (Wei and Ma, 2022) [63]. (3) Although the industrial enterprises listed in the GEM board have better financing efficiency than the enterprises in the other two boards, it is still necessary to strengthen the supervision of the GEM market, prevent various illegal acts that may occur in the financing process of enterprises, protect the legitimate rights and interests of investors, and create a good financing environment for the development of GEM listed enterprises. (4) Given the relatively backward economic development level and low degree of marketization in the central and western regions of China, enterprises in remote areas cannot form factor clusters with scale effects simply relying on the market due to the externalities of high transaction costs (Li, 2017) [57]. The resource mismatch in the central and western regions is higher than that in the eastern region, while that in the central regions is higher than that in the western region (Li, 2012) [64]. The country should accelerate the marketization reform in the central and western regions, reduce the degree of government intervention, and give certain financial preferences, such as giving preferential loan policies and simplifying the loan process, etc. In the period of rapid development of digital finance, it is necessary to promote the regional coordinated development of digital finance, give preferential support to the construction of financial infrastructure in the central and western regions, and promote the release of financing constraints of enterprises, to strengthen technological innovation of enterprises, increase their financing ability, and improve financing efficiency.

6. Conclusions, Limitations and Future Research

6.1. Conclusions

The financing efficiency of industrial listed enterprises in China is generally low, and the financing efficiency varies greatly from the nature of their property right, the listed board, and the region.

Concerning property right, the overall financing efficiency of SOEs is lower than that of NSOEs, mainly because the average fund raising efficiency of NSOEs is significantly higher than that of SOEs, the average fund using efficiency of NSOEs is greater than the fund raising efficiency, but the opposite is true for SOEs. Capital is the main resource of enterprises. The differences in ownership will cause a resource mismatch. To improve financing efficiency, SOEs should mainly improve their fund using efficiency, as opposed to NSOEs which should mainly improve their fund raising efficiency.

In terms of the listed board, the overall financing efficiency of the main-board-listed enterprises is the lowest, and that of GEM board listed enterprises is the highest; the most obvious gap lies in the fund using efficiency, but this gap is gradually narrowing. The average fund raising efficiency of the first stage of the main-board-listed enterprises is greater than the average fund using efficiency; on the contrary, the fund using efficiency of the SME board enterprises and GEM enterprises is greater than the fund raising efficiency. This indicates that the main-board-listed enterprises still have more advantages in financing, while the SMEs and GEM enterprises have stronger vitality.

In regard to the region, the overall financing efficiency of China’s industrial enterprises has obvious regional characteristics. The difference in the fund raising efficiency in each region is not large, while the difference in the fund using efficiency is obvious. On the whole, there is a pattern of high in the eastern region and low in the western region, and the financing efficiency of listed enterprises in each region is extremely unbalanced.

Based on the DNSBM model, this study not only analyzed the overall financing efficiency, but also pointed out the impact of fund raising efficiency and fund using efficiency on the overall financing efficiency, and carried out analyses based on the different natures of the enterprises. This helps enterprises to establish a clearer direction, and aids the government in formulating relevant policies to improve the financing efficiency and help enterprises achieve high-quality development.

6.2. Limitations and Further Research

Although this paper has reached some valuable conclusions, there are still some shortcomings in this paper due to the limitations of the authors’ knowledge and other factors: (1) This paper uses the financial expenses to measure the debt cost. However, there will be a certain deviation, because the financial cost in the financial statements is the balance after deducting interest income, etc., and does not fully reflect the interest expenses of debt financing. For the measurement of equity cost, this paper utilizes the capital asset pricing model (CAPM). Although a lot of research in China also adopts the CAPM model to measure equity cost, its adaptability is still controversial. Therefore, in future research, the selection of financing cost indicators can be continuously improved. (2) The study is limited to China only. Because China has a large number of SOEs, the securities markets and regional characteristics are quite different from other countries, the international applicability of the conclusions on the financing efficiency of industrial enterprises with these characteristics is limited, but the results of the financing efficiency of NSOEs and research methods have certain reference significance for other countries and researchers. It is necessary to test the validity of the findings by extending the study to other countries with similar socioeconomic characteristics. (3) The research time span does not include the most recent 5 years, as well as the impact of COVID-19 on enterprises’ financing efficiency; these are all issues that our future research will continue to explore.