Regional Economic Environment, Compensation Marketization and Financing Platform Performance—An Empirical Analysis Based on Data of China

Based on the basic theory of labor market and compensation incentive, this paper deduces the impact of compensation marketization on the performance of financing platform, discusses the differential effect of compensation marketization on the performance of financing platforms under different regional conditions, and puts forward the research hypothesis. This paper collected the financial data of 115 financing platforms in Jiangsu and Anhui provinces of China from 2005 to 2020 for empirical analysis. Our research finds that compensation marketization can improve the performance of financing platforms, and this effect will be larger if the financing platform is in a more developed region than in a less developed region. After further research we find that compensation marketization promotes performance by increasing capital turnover and reducing cost stickiness. Also executives’ political background may lead to a higher effect of compensation marketization on financing platform performance.


Introduction
In China, there are different kinds of local government financing platforms (hereafter referred to as "financing platform") founded by different levels of government (i.e. provincial, municipal and county governments). These financing platforms are state owned firms, their main function is to produce public goods such as infrastructure, public facilities, sewage disposal, etc. Because of their huge amount of investment, financing platforms make great contributions to promoting urban construction and economic development of China (Morten, 2010). However, with their development, more and more funds are needed to invest in public projects, these financing platforms have to raise funds through bank loan or bond issuance, while these funds cannot be recycled back quickly, this leads to higher financial risk to financing platforms and local government. Therefore, the transformation of financing platforms has become an urgent problem in China. It is a common means for Chinese state-owned enterprises to transform from state administrative mode to modern corporate governance (Cai, 2014). Concerning research on the transformation of financing platforms, scholars mainly focus on the financial risks of platforms (Liu & Li, 2013;Zhao, 2018;Wei, et al., 2019), risk prevention and control (Zhang et al., 2019). A few authors discuss the impact of corporate governance improvement on the performance of financing platforms from the perspective of equity diversification (Li & Wang, 2014). Governance structure is the vitality of organization operation, as an important part of corporate governance, compensation marketization has been gradually implemented in some financing platforms in China. Here, we refer to compensation marketization as the similarity of employees' compensation system to that of most private firms. But there are few researches study whether such reform can achieve corresponding results. This paper collect the data of some financing platforms in Jiangsu province and Anhui Province of China that publicly issued bonds from 2005 to 2020, to explore whether the compensation marketization can improve the operating performance of financing platforms, and study whether the improvement of operating performance may be different with regional difference.
As public goods provider, financing platforms and other large state-owned enterprises in China such as tobacco, electric power, salt and other such producer, has the natural monopoly. The monopoly ensures the national control to these industries, while it also weakens the effect which product market competition exert on corporate governance, thus may causes low efficiency. Therefore, it is of great significance to improve the corporate governance mechanism by using compensation marketization while keeping state control. In terms of academic value, the contribution of this paper is as follows: (1) we analyze the incentive effect of compensation according to the basic principle of "distribution according to work", that is, in a perfect competition labor market, each job has its equilibrium price. Only when the compensation of the job equals its equilibrium price, the firm gets the highest efficiency. Thus we expand the traditional research paradigm which is limited to study the compensation-performance sensitivity to measure the incentive effect of compensation. (2) Combining the labor market and product market competition, comprehensively analyze the role of market competition in the corporate governance of financing platform and other state-owned/holding companies which have certain product market monopoly. (3) This paper found that the implement of the compensation marketization of financing platforms contributes to the performance improvement, which provides a different view from the existing research conclusion, which do not find significant evidence to confirm the positive relationship between compensation incentive and firm performance under low market competition conditions. (4) This paper pays attention to the impact of compensation marketization on financing platform performance under different regional conditions, and provides appropriate advice for the transformation policies of financing platforms in different regions in China.

Literature Review
In China, financing platforms are investment and financing companies established by the local governments by investing land, equity and other assets (Zhou, 2011). And they undertake the function of financing urban infrastructure construction and operation for the local governments. With the rapid expansion of financing platform, financial risk of these platforms and the government debt risk gradually revealed. To prevent risks, the Chinese central government has strengthened management of local financing platforms (Cao et al., 2020). In 2014, the new Budget Law of China stripped part of the functions of financing platforms for governments financing (Jiang & Yu, 2020). The market-oriented transformation of financing platforms has become the main direction of development. Scholars have made various studies on this subject. Many authors studied the advantages and disadvantages of financing platforms marketization from different aspects such as introducing private capital (Zhang & Ouyang, 2017;Hu, 2017) or improving financing platforms' corporate governance (Li & Zhang, 2019). Also, some researchers suggest to reform the relationship between government and financing platforms (Mao et al., 2018), how to control financial risk of financial platforms (Wen, 2013) and to introduce more private products in financial platforms (Wang & Li, 2020).
In the process of market-oriented reform of financing platform, compensation marketization is the key (Du, 2018). How to take advantage of the incentive role of compensation is the main task of reform of SOEs (Song & Meng, 2012). The research about incentive effect of compensation includes the following: (1) the relation between absolute value of compensation and performance relationship. Most studies show that under certain degree of product market competition conditions, compensation incentive is conducive to improving the company's performance, while when the product market competition is low, the relationship between compensation and performance is not obvious (Liu et al., 2007). (2) The impact of relative compensation incentives on corporate performance. On one hand, the increase of internal salary gap in a certain range has a positive impact on the improvement of enterprise performance, when the internal pay gap exceeds a certain extent, it will adversely affect the performance (Jed, 2006). On the other hand, there is a significant compensation reference point effect on executive compensation, that is, if employees find that their salary is below the market salary benchmark, it will affect their work enthusiasm, thus leading to low company performance (Luo et al., 2019;Wang & Geng, 2020). As to the compensation incentive of financing platforms, due to the low correlation between executive efforts and performance, executive compensation is weakly related to executive and financing platform performance (Chen et al., 2018).
In addition to internal factors and industry factors, the regional economic environment of an enterprise also has an effect on enterprise's performance. Due to the differences in resource endowment and regional economic policies, there are obvious differences in economic development levels among different regions in China (Shao & Yang, 2010), The GDP scale of the eastern region is significantly higher than that of the central and western regions. Regional differences have an important impact on regional enterprises in capital, labor, infrastructure, technology supply, financial environment and many other factors. In terms of geographical location and investment environment, a firm of eastern coastal regions of China have obvious advantages over that of central and western region and can easily obtain capital and other resources (Mi, 2008). Firms in eastern region can also get more technological resources than that of other regions (Fang, 2020). It is clear that regional economic environment is an important exogenous factor influencing enterprise development. Regional factors such as capital, labor, technology and system will affect enterprise cost and strategic choice, and then affecting enterprise performance (Wu et al., 2008;Xu et al., 2011).
Although much progress has been made as recited above, there are still areas worthy of further research:(1) most research on the role of compensation incentive in state-owned enterprises which are monopoly in the product markets mainly discusses the impact of absolute compensation, e.g. the sensitivity between compensation and performance, and ignores the impact of relative compensation, especially compensation marketization on enterprise performance. In fact, with the advancement of the salary system reform of state-owned enterprises in China, the compensation marketization in state-owned enterprises such as financing platforms has gradually increased. According to the theory of compensation reference point effect, this improvement in compensation marketization will have an impact on performance, but there are few researches study the impact.
(2) Most research on the compensation incentive of state-owned enterprises mainly focuses on the inefficiency of the compensation incentive caused by the weakened competition in the product market, while the impact of labor market competition on SOE compensation incentives is not fully considered. From a realistic point of view, with the gradual improvement of China's market economy system, labor liquidity is enhanced. Competition in the labor market leads to the increased liquidity of enterprise employees. Therefore, even in state-owned enterprises, because employees have certain market liquidity, enterprises need to form an employment relationship with employees according to the market compensation. Thus the incentive effect of the compensation marketization is formed in SOEs. (3) Most empirical studies choose one platform or platforms in a certain region as research sample. But in China, there are a large number of financing platforms in different regions, so how regional factors affect financing platforms is still unknown to researchers.

1) Background of compensation marketization reform of financing platform
Most of the products (i.e., highway, railway, airport, etc) produced by the financing platforms are public goods, and each financing platform is the sole producer of these products, so financing platforms are monopoly companies in their product markets. Because financing platforms are administered by the local governments, their compensation systems are different from that of other private firms. The employees' compensation of these platforms are set according to civil servant salary system, they do not change with financing platform's performances, which makes the compensation of the financing platform rigid. Thus, the compensation marketization of financing platforms is low compared to private firms, and this is not conducive to motivating employees. With the reform of state-owned enterprises' compensation system and the implementation of financing platform transformation, some financing platforms began to reform the compensation system. The compensation marketization of state-owned enterprises is a relatively long process. Before 1978, most enterprises in China are state-owned enterprises, compensations of these enterprises are set by national labor rules, each enterprise can only pay employees' compensation according to these rules. In 1979, the Regulations on Profit Retention in State-owned Enterprises proposed to draw reward funds from profit retention to link employees' income with their performance. In 2000, the National Wage Conference put forward the model of "market mechanism deciding, enterprise independent distributing, government supervising and guiding", which made it clear that the market plays a decisive role in compensation distribution. The Opinions of the State Council on Reforming the Wage Determination Mechanism of soes released in May 2018 further pointed out that the decisive role of the market in the wage distribution of SOEs should be given full play. And this policy plays an important role in promoting the market-oriented reform of the compensation of financing platforms. From the perspective of financing platform enterprises, in order to improve the efficiency of compensation system, some platform companies began to explore the compensation system reform to match compensation with the performance and market value of financing platforms (Ding, 2018). However, due to governmental management system, the compensation marketization is still low compared to that of local private enterprises.
For most of state-owned or state-controlled enterprises in China, the market-oriented reform of compensation has achieved good results. Most studies have confirmed that the moderate increase of executive and employee compensation, appropriate performance assessment and incentive system can effectively improve SOEs' performance (Lu et al., 2011;Xia & Dong, 2014;Cai et al., 2018). However, in the case of low market competition, the correlation between enterprise performance and management incentive is relatively low (Liu et ijef.ccsenet.org International Journal of Economics and Finance Vol. 14, No.4;. According to the above research conclusions, the increase of absolute compensation will not have an obvious incentive effect on the financing platforms. However, considering that the promotion of compensation marketization reform of financing platform may also have an incentive effect from the perspective of compensation fairness, we analyze its impact on the performance of financing platform.

2) Compensation marketization and enterprise efficiency
Based on the basic theory of labor economics, using the analysis framework of competition and compensation in labor market, we analyze the relationship between compensation and performance in a firm. In the labor market, the compensation price of a certain wokload is determined by its supply and demand. Referring to the analysis model of Borjas (2018), we explore the formation of compensation and workload in the labor market. The details are shown in Figure 1 below:

Figure 1. Diagram of compensation formation under full competition labor market
The left section of Figure 1 depicts the relation between compensation and labor supply, the relation between compensation and labor demand, and the ultimate compensation in the labor market equilibrium under different workload post levels 1, 2, and 3. S1, S2 and S3 respectively represent the changes in the labor supply curve when the job workload increases from low to high. As the job workload increases, the curve slope becomes smaller, and the labor supply decreases. D1, D2 and D3 represent changes in the labor demand curve when the workload increases from low to high. From D1, D2 and D3 the curves go upward, it means that as the job workload increases, companies are willing to pay more compensation. The intersection of D1 and S1, D2 and S2, and D3 and S3 are P1, P2 and P3, respectively, indicating the market compensation price for different job workload when the market reaches the supply and demand equilibrium. When we combine the market equilibrium prices of different workload levels into a curve, we get the market compensation curve in the right section of Figure 1. Market compensation curve shows that in fully competitive labor markets, compensation increases as job complexity or workload increases. The shaded part with the solid lines above the curve is the employee benefit area, where employees are willing to work because the compensation exceeds the requirements by the employee workload. From the point of view of firms, they need to hire at the minimum cost. As the job workload and complexity increase, the compensation that firms are willing to pay increases. Under the same circumstances, the lower the compensation, the more inclined the firm is to hire. Thus forms the shaded portion formed by the dashed line below the curve in the right section of Figure 1, e.g. firm benefit area. In the fully competitive labor market, the intersections of the two shadows form a market compensation curve. Employees are paid according to the market salary curve and provide the corresponding quality of work, while the enterprise employs employees according to the market compensation, and obtains the corresponding quality of employee work. Due to the existence of market compensation curve, the difference between actual compensation and market compensation curve becomes the factor affecting enterprise performance. As shown in the right section of Figure 1, assume Point A, B and C represent three different employment relationship between enterprises and employees. Considering point A, employee workload is equal to that of point C, but the firm pays more compensation; concerning point B, the employee does more than that of point C, but gets the same as point C, the work enthusiasm of employee is reduced. Only when at point C, the relationship between compensation and workload is in the compensation market curve. Compensation and workload reaches market equilibrium. Compensation efficiency is the highest, enterprises achieves the maximum efficiency. Thus we get the following inference: Inference 1: The closer the job compensation is to the market compensation curve, the higher the enterprise efficiency.
In practice, the labor market can not always be in a completely competitive state，the actual relationship between employee workload and compensation may not be on the market compensation curve, so there is a certain distance from the market compensation curve. The size of the distance marks the degree of marketization of enterprise compensation. The distance between points A and C, B and C, D and C in Figure 1 indicate the different degree of marketization of employee compensation. The workload at point D is the same as that at point A and C, but the distance between A and C is larger than the distance between D and C, it shows that the compensation marketization degree of the point D is greater than the point A. According to Inference 1, it is easy to see that the greater the marketization of compensation, the closer the employee salary is to the compensation market curve, and the higher the efficiency of the enterprise.
From the perspective of the actual compensation system of financing platforms, due to the different measures taken by different financing platforms in compensation reform, the gap between actual compensation and market compensation is different among different financing platforms, that is to say, the degree of marketization of compensation is different among different financing platforms. According to Inference 1, we propose following hypothesis: Hypothesis 1: Compensation marketization can improve the performance of financing platforms.

Moderating Effect of Regional Economic Environment on the Relationship between Compensation Marketization and Performance of Financing Platforms
China has a vast territory and has different economic environment among different regions. Usually in regions with better economic development, the number of enterprises and labor reserves are more, and their labor market is more developed, thus it is easy to form a relatively stable labor prices (Yang & Xu, 2007;Wang & Liu, 2016), and the market compensation curve has clear boundaries. In less developed regions, the number of enterprises and labor reserves are relatively small, and the labor market is relatively depressed, so it is not easy to form a stable labor market price.
We still use the workload and labor market price diagram to analyze the impact of different regional economic environments on the relationship between compensation marketization and enterprise performance. When a region where the firm locates in is economically developed, the labor market is relatively developed and the market labor price is relatively stable, thus forms a clear market compensation curve as shown in Figure 1. In this case, the higher the degree of compensation marketization, the higher the efficiency of compensation incentive, and the enterprise performance is more improved. When a firm is in a less developed region, the relationship between market price of compensation and workload is shown in Figure 2: As shown in Figure 2 above, in the labor market of the underdeveloped regions, because there are not enough enterprises and employees in the market, the market in most cases can not achieve the equilibrium between supply and demand and form a stable market compensation curve. Companies will limit the maximum acceptable compensation for each workload of different jobs, while employees will limit the minimum acceptable compensation. When the compensation value of the firm demand curve is higher than the compensation value of the employee supply curve, the firm and the employee may reach an employment contract, thus forms an employment agreement region, as the shaded part shown in Figure 2. Once the compensation value of the employee supply curve exceeds the compensation value of the firm demand curve, the firm and the employee cannot reach an employment agreement. As shown in Figure 2 above, the firm employs employees to work in L1 position. Since the labor market does not achieve full market competition, employee's compensation may be at points B, A and C. When the workload of a job exceeds L2, no employment agreement cannot be reached between the employee and the firm.  Vol. 14, No.4; In figure 2 above, the area of the employment agreement region is affected by the development of the labor market. The weaker the labor market, the lower its supply and demand. The maximum compensation firms pay for employees will increase, and the minimum compensation employees are willing to accept will lower. This increases the width and area of the employment agreement region. On the other hand, when the labor market becomes more developed, the width of the employment agreement area is reduced, and the area decreases. When the market reaches full competition, the employment agreement region is reduced to a curve, e.g. market compensation curve, its area is zero.
The development of the labor market has an impact on the efficiency of enterprise compensation incentive. In the developed labor market, there is a clear market compensation curve. Employees provide labor according to the market compensation curve and receive market compensation, the compensation incentive efficiency is higher.
In the underdeveloped labor market, there is no clear compensation market curve. Firms compensation lacks a clear reference point from the market, unfair compensation phenomenon such as different compensation for the same work (as Points A, B, and C shown in Figure 2) or different work for same compensation will be more likely to appear, thus reduce the efficiency of compensation incentives. According to all these, we propose the following hypothesis: Hypothesis 2: In regions with better economic environment, the marketization of compensation plays a more significant role in improving the performance of financing platforms.

Data Sources
Considering the large number of financing platforms across the whole China, and most of the financing platforms are unlisted companies, the financial report is not public, it is difficult to obtain all samples of financing platforms across the country. We select the financing platforms of Jiangsu Province and Anhui Province from 2005 to 2020 as the research sample, and obtain the annual reports and bond prospectuses of these financing platforms from China Money Network and Bond Information Network. Though Jiangsu province and Anhui Province are both in the east of China, the economic environment of the two provinces is different, so we can analyze how regional differences affect platforms' performances. We collected a total of 115 financing platforms in Jiangsu and Anhui Province, 1000 financial reports from 2005 to 2020. The compensation marketization data for matching each financing platform is obtained from the annual reports of the local listed companies of the financing platform, and the GDP growth data of the location of each financing platform is obtained from the statistical reports of national economic and social development published by the provinces and counties where the financing platform is located. We delete outliers and winsorize the data at 0.01. We use Stata15 to analyze the data.

2) Dependent variable
Financing platform performance (Realroa): In this paper, return on total assets is used to measure the performance of a financing platform. Since the financing platforms mainly produce public products with low profitability, local government will give certain fiscal subsidies to these platforms, which will have a certain impact on its performance, so we have to exclude fiscal subsidies in the calculation of Realroa. Here, the formula of return on total assets is: Realroa=(annual net profit -annual government subsidy)/total assets at the end of year

3) Independent variables
Explanatory variable P in Model (1) is used to measure the compensation marketization. We refer to the concept Sun et al. (2017) adopted when studying the positive and negative aspects of the marketization of SOE executive compensation. Since private enterprises are more market-oriented, the smaller the compensation gap between SOEs and private enterprises, the SOE executive compensation is more in line with the market, thus the compensation marketization is higher. As most of the financing platforms are unlisted companies, we cannot obtain the specific compensation data of their executives and employees. Therefore, the overall compensation of the platform company is measured by the ratio of "cash paid to and for employees" in the Cash Flow Statement ijef.ccsenet.org International Journal of Economics and Finance Vol. 14, No.4; of the platform company by the Operating Revenue in the Income Statement. The absolute value of the difference between the data of the platform company and the corresponding data of the listed company where the platform company is located is taken as an indicator to measure the marketization level of compensation. The larger the value of this index is, the lower the marketization level of compensation of the platform company is.
The definition and measurement of all the variables used in the model are shown in Table 1: According to annual GDP growth percentage in year t published by local government Yeart The year in which financial statements are made

Descriptive Statistical Results
The descriptive statistical results of the main variables are shown in Table 2. In Panel A, the average value of Realroa is only 0.008, much lower than the average value of listed companies and ordinary enterprises. This may be caused by the low profit margin of the financing platform's products. On the other hand, it may also be caused by the non-existence of real owners, which leads to low corporate governance efficiency and agency problems. The mean value of P is 0.1, indicating that the difference between the salary level of financing platform companies and that of general listed companies is not very large. However, the large standard deviation indicates that the salary marketization of different financing platform companies still varies greatly. The mean of Lever is 0.541, indicating that the overall debt level is not high. The mean value of Ind is 6.252, but the standard deviation is 3.910, indicating that the diversification degree of financing platform industry varies greatly. The mean value of Zonghe is 0.748, indicating that most of the samples are municipal project producers. The mean value of Shengji, Xianqu and Shiji was 0.101, 0.144 and 0.756 respectively, indicating that municipal financing platforms were in the majority and provincial financing platforms were the least. The mean value of Size is 24.06, indicating that the average asset size of financing platform is basically similar to that of listed companies in China. The mean margin was 0.155, indicating low profitability. The mean value of Top1 reached 0.992, indicating that the vast majority of financing platforms are absolutely controlled by state. The mean value of Gdp is 9.155, while the standard deviation is 3.127, the minimum value is -1.7, and the maximum value is 18.1, indicating that the GDP growth rate of cities in Jiangsu province and Anhui Province is generally large, but there is a big difference.
Referring to the statistics of provincial samples, in Panel B, Realroa of Jiangsu province is slightly lower than that of Anhui Province, which may be mainly due to the low degree of diversification of industry and property rights in Jiangsu Province platforms (Ind is lower than Anhui Province, Top1 is higher than Anhui Province). Due to the low degree of industrial diversification, margin in Jiangsu is lower than Anhui Province. The mean of P of Jiangsu Province is 0.097, lower than that of Anhui Province (0.103), indicating that compensation marketization of financing platforms in Anhui Province is lower than that in Jiangsu Province.  Table 3 reports the correlation coefficient between the main variables, and we can see that P is significantly negatively correlated with Realroa (The level of significance is lower than 0.01). P reflects the absolute value of the difference between the compensation level of financing platforms and general listed companies. The smaller the P, the higher the compensation marketization of financing platform. The significant negative correlation between P and Realroa gives the preliminary evidence of hypothesis 1.

Multiple Regression Results
The results of multiple regression analysis are shown in Table 4. P in Column (1) and (2) is significantly negatively correlated with Realroa(The level of significance is lower than 0.01), indicating that compensation marketization level is significantly positively correlated with performance. In other words, compensation  (1) and (2) (The level of significance is lower than 0.01), indicating that the higher compensation to employees is conducive to encouraging employees. Even in the public goods industry which has low market competition, compensation improvements still boost the company's performance. The result does not agree with current studies. The possible reason may be that in firms such as financing platforms which face low product market competition, though raising compensation cannot lead to higher market revenue, but can promote employees to improve their work through other channels, thus improve firm performance. The significant negative correlation between Lever and Realroa indicates that the financing platform with a higher debt ratio has a lower profitability due to its high debt capital. The significant positive correlation between Zonghe and Realroa indicates that the products of the comprehensive financing platform are closer to the market demand and the business performance is better. All these results are consistent with economic theory.
Column (2) shows the regression results after the addition of Provin. The cross-product P_Provin (P×Provin) is significantly negative correlated with Realroa (The level of significance is lower than 0.01), this means that there are differences in the effect of compensation marketization on improving performance in different regions, and this effect is more obvious in Jiangsu Province. According to the sub-sample regression results, as shown in Column (3), P in Jiangsu province is significantly negatively correlated with Realroa (The level of significance is lower than 0.01), it means that compensation marketization can improve operating performance of financing platform enterprises in Jiangsu Province. In Column (4), the relation of P and Realroa of performance in Anhui province are not significant, which may be caused by the low marketization level of compensation of financing platforms in Anhui Province. All these results confirm Hypothesis 2.

Robustness Test
In order to ensure the reliability of our research, we carry out robustness tests: (1) The robustness test is carried out by changing the dependent variable: we use Roe (net profit of the platform/net equity of the platform) to replace Realroa to measure platforms' performance.
(2) In order to avoid endogeneity problems, we delayed the dependent variable Realroa for a year in the Model (1) and do regression again.
(3) Propensity score matching (PSM): The propensity score matching method was used to construct the samples of experimental group and control group for PSM matching. (4) Instrumental variable method: we delayed variable P of sample as the instrumental variable. Through all these tests, the results are basically consistent with the original conclusions. See Table 5 for details of robustness test.

Further Research
Though we have proved that compensation marketization improves the performance of financing platforms in former section, the mechanism of this effect needs to be further confirmed. In addition, compensation marketization is a reform project for financing platforms, and its promotion process is bound to encounter some resistance. What mechanism financing platforms use to guarantee the promotion of compensation marketization is also worthy of in-depth study. We conduct further research on the above issues.

1) Asset turnover perspective
The essence of compensation marketization is to make employees get compensation relative to the quality of their jobs. Such reform measures encourage employees to improve the quality of work and reduce the production ijef.ccsenet.org International Journal of Economics and Finance Vol. 14, No.4; and operation costs of firms, so as to achieve the purpose of improving performance. Considering that financing platforms are capital-intensive enterprises and slow in capital turnover due to long infrastructure construction cycle, capital cost becomes an important factor affecting their performance. In order to improve the performance of financing platform, it is necessary to improve asset turnover and reduce capital occupation, thus reducing the cost of capital. On the one hand, compensation marketization strengthens the internal management of platform companies, makes the management more prudent and enhance capital turnover. On the other hand, compensation marketization improves employees' work enthusiasm so as to improve the efficiency of asset use in the production process, reduce the total amount of capital occupied and improve the performance of the financing platform. Referring to the mediation effect testing procedure proposed by Wen and Ye (2014), the regression model is established as follows to test the mechanism of asset turnover as an intermediary variable: The empirical results are shown in Table 6. We use the sequential test method. First, the estimation coefficient of P in model (2) is significantly negative, indicating that the total effect of P on platforms performance is significantly negative. In the second step of partial mediation test, P coefficients in Model (3) and Model (4) are significant, indicating that partial mediation effect is established. And the absolute value of the P coefficient in model (4) is less than that in model (2), indicating that the coefficient difference test has been passed and the mediation effect is confirmed. That is, compensation marketization can improve the operating performance of financing platforms by improving asset turnover and reducing capital cost.

2) Cost stickiness perspective
Before the reform of the compensation system, due to the difference between the compensation and the market level, and the imperfection of internal control system, other forms of income, such as allowances and on-the-job consumption, have become the way for employees to increase their own income. This not only reduces the performance of the financing platforms, but also increases the expense stickiness. The implementation of compensation marketization makes the enterprise employees get a more fair compensation, thus lower the motivation for employees to get income from other sources. Meanwhile, as the improvement of compensation system such as "Link Compensation with Efficiency" begins to work, employees are motivated to reduce improper expenses, to improve performance and personal compensation. Therefore, we infer that compensation marketization can improve operating performance by reducing platforms' cost stickiness. In order to confirm the inference, we refer to the measurement method of Wang and Xi (2011), set variable Fynx to measure cost stickiness of financing platforms. Its value is measured as: Fynx=ln(overhead of year t/overhead of year t-1)-( operating income of year t/operating income of year t-1).
We establish a regression model to test the effect mechanism of cost stickiness as an intermediary variable: The regression results are shown in Table 7. The sequential test method is adopted. The result confirms that compensation marketization can improve the operating performance of financing platforms by reducing cost stickiness.

Analysis of the Effect of Political Background on Compensation Marketization and Financing Platform Performance
At the early age, financing platform adopts the salary system of civil servant model, thus employee compensation system has little relation with financing platform performance (Jin, 2016). The compensation system reduces the risk of salary being affected by market performance to a certain extent, and makes the correlation between employee compensation and operating performance of financing platform low. On the one hand, the implementation of compensation marketization may provide a certain space for the increase of employees' compensation, but on the other hand, it leads to the increase of the uncertainty of employees' compensation and the enhancement of performance constraints. Therefore, the implementation of this policy will cause great resistance within financing platforms. Most of the financing platforms are wholly state-owned enterprises. Due to the absence of owners (Note 2), the implementation of reform measures lacks the support of the authority of the equity owner, and the insiders (management) become the ultimate guarantee for the implementation of policies (Zhang, 2005). The political background of the management of state-owned enterprises is an important support for the authority of the management. The stronger the political background, the richer the political resources they can control, and the more capable they are to use various resources in the organization and implement systems and policies. However, under the current official evaluation mechanism in China, the superior government evaluates local government officials mainly by assessing the local infrastructure construction, industry development and social and economic development during the term of office of local officials. Such political achievements (i.e., local infrastructure construction, industry development and social and economic development, etc.) greatly affect the promotion and political reputation of local officials. The financing platform undertakes the function of investment and financing for the government. The directors and senior executives of the platforms are appointed by the government, so they have strong motivation to fulfill the tasks required by the local government in their work. To implement the reform policy of market-oriented compensation of financing platforms, senior executives with political background will mobilize their resources to overcome the resistance to improve the performance of financing platforms for the purpose of career assessment or promotion. Therefore, it is obvious that the political background of senior executives in financing platform companies will promote the marketization of compensation to improve the operating performance of financing platform.
In order to verify the inference, we collected the data of the top three executives of 108 financing platforms, set variable "Govern" to measure the total of political background according to the classification standard of civil servant posts according to Wang and Rong (2018) and National Civil Servant Law (2005) (Note 3). We add Govern and its cross-product with P (Govern×P) into regression analysis model (1). The results of descriptive statistics and regression analysis on Govern were shown in Table 8 and Table 9.
It can be seen from Table 8 that there is little difference between the mean values of executive political background in Jiangsu province and Anhui Province on the whole, but the standard deviation of both provinces is relatively large, maybe there is a large difference in executive political background among financing platforms of different levels. The political background of executive of provincial and municipal financing platforms is stronger than that of district and county-level financing platforms. In Column (1) of Table 9, there is a significant positive correlation between the political background of senior executives and the dependent variable Realroa, which may due to the stronger the political background in an SOE, the more external resources management can obtain for its operation, thus boosting its performance. In column (2), P_Govern, the cross term of P and Govern, is significantly negatively correlated with Realroa, which proves that Govern (executive political background) can significantly promote the improvement of compensation marketization on business performance. This shows that the political background of senior executives of financing platforms will promote the improvement of compensation marketization on business performance, confirming our inference. The sub-sample test results of Jiangsu province and Anhui Province are shown in Columns (3) and (4). P_Govern in column (3) is significantly negatively correlated with Realroa, which proves that the political background of senior executives in Jiangsu Province can play a role, while it is not significant in the data of Anhui Province, which may be caused by the low compensation marketization level in Anhui Province.

Research Conclusion and Implications
Based on the data of financing platforms in Jiangsu and Anhui provinces of China, we conduct theoretical analysis and empirical test, and draws the following conclusions: (1) Implementing compensation marketization in financing platform enterprises is helpful to improve enterprise performance.
(2) Compensation marketization plays a more significant role in improving the performance of financing platforms in more developed region (Jiangsu Province) than in less developed region (Anhui Province). In addition, we find that compensation enhancement in the financing platform could improve its performance. Further research shows that compenstion marketization can improve the performance of financing platforms by improving asset turnover, reducing cost of capital and expense stickiness. The political background of senior executives of financing platforms is conducive to the promotion of compensation marketization to platforms performance.
Based on the above research conclusions, this paper has the following implications. First, in view of the promotion effect of compensation marketization on the performance of financing platforms, financing platforms at all levels can take measures to promote compensation marketization to further improve their performance. Especially in developed regions with relatively good economic development, reform can be accelerated to obtain greater returns. Second, because compensation increasing is helpful to the increasing of financing platform's performance, the financing platform can appropriately improve the intensity of compensation incentive to get better performance. Third, local governments should introduce more effective policies, to promote compensation marketization in financing platforms, and not give much burden on management of platforms. Fourth, considering the role of compensation marketization on improving the performance of the financing platform, it is suggested that China's large state-owned specialized enterprises such as telecom, oil and salt industry promote the process of compensation marketization, to improve governance and performance.
This paper focuses on the positive effect of compensation marketization on the performance of financing platforms and takes note of the different impact of different regional economic development levels. However, due to data limitations, we fail to separate the marketization of executive compensation and employee compensation, and respectively discuss their impact on the performance of financing platforms. Subsequent researches can further study the possible differential impact of the marketization of different employees' compensation.