Compensation regulation and political promotion of executives of state-owned enterprises——Quasi-natural experiment based on the reform of compensation regulation in China

ABSTRACT Using a sample of listed state-owned enterprises in China from 2013 to 2018, this paper focuses on the impacts of compensation regulation policy on the changes of incentive mode for SOE executives. We find that after the policy, the compensation incentive channel for highly-paid executives has been replaced by political promotion as an alternative incentive mechanism. Further analysis shows that ingratiatory behaviour has become the main channel for SOE executives to seek promotion after the policy. Our results provide support for the relationship hypothesis relating to the political promotion of SOE executives, and such effects are more pronounced within younger SOE executives with lower enterprise level and marketisation degree. Finally, we find that the existence of alternative benefit acquisition channels such as equity incentive, perquisite consumption and the degree of mixed ownership reform of SOE has weakened the relationship between compensation regulation and political promotion of SOE executives.


Introduction
Traditional incentive theory points out that there are mainly two executive incentive mechanisms. One is the material incentive mechanism represented by compensation incentive; the other is the promotion incentive. However, due to the long-term, uncontrollable and space-limited 1 nature of promotion, executives will pursue material incentive (compensation) in the short term. Prior literature suggests that executives will wield their influence to seek rent which leads to morbidly excessive executive compensation that far exceeds their earning power through fair negotiation (Hambrick & Finkelstein, 1995;Liu & Zhao, 2008). Abnormally high executive compensation has attracted wide attention since 1980s. During the 2008 global financial crisis, the 'sky-high' executive compensation of the U.S. financial sector, especially investment banks, 2 has brought this issue under the spotlight. In China, public has cast doubts on the executive compensation with the compensation gap between executives and employees enlarged. This is particularly serious with SOEs in China (Quan et al., 2010). In addition to the aforementioned factors, such as management power, 3 government support for SOEs also facilitates executives to pursue excessive compensation, thus making SOE executive compensation deviate from the optimal level (Chen et al., 2005). A direct consequence of such deviation is the 'skyhigh' compensation of SOE executives.
In order to maintain social fairness, various provisions have been introduced all over the world to restrict executive compensation. The U.S. government announced a compensation regulation policy on Wall Street in 2009. The Chinese government began regulating the compensation of SOE executives as early as 2002. Since then, the Chinese government has issued a series of documents to regulate the compensation of SOE executives, the strictest among which is the Compensation Regulation Policy (CRP) introduced in 2015. This policy mainly targeted the executives under the administration and appointed by the Central Enterprises (other SOEs refer to the implementation), where it was specifically regulated that, the annual basic compensation, annual performancerelated compensation and tenure-related compensation of these SOE executives should be within certain range in comparison with the average wages of the employees.
With the inception of this CRP, a drastic drop was observed in the level of compensation of SOE executives, as well as the associated incentive effect. Motivated by selfinterest, executives would then seek opportunities to compensate for the disbenefits caused by the institution. Unlike executives in non-SOEs and other countries, SOEs executives in China pursue not only compensation incentive as managers, but also political promotion as politicians. During the period of economic transformation in China, the administrative level of executive positions in SOEs still exists, and it is also quite common for SOE executives to be managed by the organisational departments of party committees (Kato & Long, 2011). Therefore, SOE executives usually have dual status as a manager and a government official (Chen et al., 2018;Yang et al., 2013). Such dual status endows SOE executives with two incentive channels. On the one hand, as managerial executives, they are incentivised by the compensation brought by their executive position (Jensen, 1986;Yang et al., 2013). On the other hand, as government officials, their appointments entail a strong administrative colour which makes political promotion critically important to their utility function. Political promotion is commonly characterised as long-term, uncontrollable and space-limited. Contrastingly, compensation incentive is considered as brisk and attainable. Therefore, before the CRP, compensation incentive of SOE executives weakened their motivation to pursue political promotion, and the inception of the CRP has strengthened the motivation to pursue 2 According to SIFMA, the total salaries paid to CEO of the five largest investment banks were still as high as $3.1 billion, despite three of them were quarterly losses in 2007. 3 With the deepening of China's market-oriented reform, senior executives of state-owned enterprises have gradually gained autonomy in production, investment, financing and personnel (Tenve et al., 2002), and the executive power is increasing (Lu, 2007). In addition, the reform process of SOEs is essentially the process of the formation and growth of the power of SOE executives (Quan et al., 2010).
political promotion of the SOE executives with the reduction of their excessive compensation and decrease of the proportion of compensation incentive in their utility function.
Existing research on the relationship between different incentive modes mainly focus on the material incentive modes, few research has explored the relationship between material and promotion incentives. Prior studies on different material incentives, providing evidence that with the reduction of compensation, SOE executives have resorted to perquisite consumption (Chen et al., 2005;Liu et al., 2009) and corruption Xu & Liu, 2013) as alternatives. To the best of our knowledge, no study has explored the impact of the CRP on SOE executives' motivation to pursue political promotion. Further, with the policy restricting the perquisite consumption behaviours of SOE executives, the new round of SOE reform, the corruption crackdown from the central government, etc. and the incentive effect of perquisite consumption and corruption minimised, we anticipate that, after the inception of the CRP, SOE executives will pay more acute attention to their political future and promotion incentive will become a substitute or supplement to the compensation incentive.
We further explore the pathway through which SOE executives get promoted after the CRP. Li and Walder (2001) argue that under the socialist system, there are two pathways for officials to get promotion. One is the professional pathway, where promotion is directly related with professional competency. The other is the political pathway, which emphasises political loyalty. Correspondingly, there are also two pathways for SOE executives to get promotion. First, the performance hypothesis suggests that SOE executives would improve corporate performance as demonstration of their competence to their superiors to seek promotion Feng & Johansson, 2017). Second, the relationship hypothesis posits that most SOE executives would actively bear policy burdens, prioritising government agenda, to seek promotion (Huang, 2003;Zhou, 2007). With theoretical support wellgrounded in prior literature for both pathways, we set out to examine which pathways SOE executives can get promoted after the CRP, and this is of great significance to shed light on the behaviour of SOE executives and offer insights to future regulations.
Using a sample of listed state-owned enterprises in China from 2013 to 2018, this paper focuses on the impacts of compensation regulation on the change of incentive mode for SOE executives. We find that after the inception of the CRP, the compensation incentive channel for highly-paid executives has been encumbered and has been replaced by political promotion as an alternative incentive mechanism. Further analysis shows that ingratiatory behaviour has become the main channel for SOE executives to seek promotion after the CRP. Our results provide support for the relationship hypothesis relating to the political promotion of SOE executives, and such effects are more prominent within younger executives with lower enterprise level and marketisation degree. Finally, we find that the existence of alternative channels such as equity incentive, perquisite consumption and the degree of mixed ownership reform of SOE has weakened the relationship between compensation regulation and political promotion of SOE executives. In conclusion, this paper reveals the important role of political promotion as an executive incentive mode and provides insightful support for further SOE reform in China.
The main innovations and contributions of this paper are as follows: (1) Existing research on the relationship between different incentive modes mainly focus on the material incentive modes, few research has explored the relationship between material and promotion incentives. This paper contributes to the exiting literature by focusing on the changes of political promotion incentive mechanism for SOE executives after the inception of the CRP. (2) To complement the existing literature on political promotion of SOE executives, where two pathways have been suggested for SOE executives to get political promotion, with theoretical support from performance hypothesis and relationship hypothesis. Focusing on the impact of the CRP on political promotion of SOE executives, this paper systematically investigates the pathways and the economic consequences of the political promotion of SOE executives. Further, this paper examines the potential institutional incentives for the political promotion of SOE executives, highlighting the theoretical underpinnings of such event.
(3) It is generally acknowledged that there are two disadvantages of government regulation. First, it is prone to be used by different interest groups in pursuit of their own agendas. Second, regulation is often divorced from reality and rigid, inducing opportunistic behaviours and distorting the spontaneous evolution of the market itself (Steven, 2002;Chen et al., 2003). This paper aims to explore the impacts of CRP on the incentive mechanisms of SOE executives, providing additional firm-level evidence to the relevant researches.

Institutional background
Unlike many countries, Chinese SOE executives not only pursue compensation incentive as managers, but also pursue political promotion as politicians. As enterprise managers, SOE executives have a strong motivation to pursue material incentives, and the government support and the absence of owners will facilitate SOE executives to pursue excessive compensation. As government officials, SOE executives have a strong desire to pursue political promotion because higher administrative levels would endow them with more resources and influence. Therefore, the incentive mechanism of SOE executives, with the synchronic existence of both compensation incentive and political promotion incentive, are different from that of other countries or non-SOEs in China.
After the reform and opening up, the reform of China's SOEs have been continuously promoted in the direction of marketisation. The traditional compensation system characterised by equalitarianism has gradually been replaced by the performance-based compensation system. However, even under the performance-based compensation system, the executive compensation of SOEs has its unique complexity. First, due to the natural connection between SOEs and the government, SOE executives can obtain policy advantages in financing, government subsidies and investment (Lu & Zhang, 2016;Pan et al., 2009). Such advantages facilitate SOE executives to pursue excessive compensation. Second, during the period of economic transformation in China, SOEs have both social goals and economic objectives. This leads to the ambiguous causal relationship between corporate performance and the executives' efforts, which makes it difficult for the government to evaluate the efforts of SOE executives (Chen et al., 2005), thus pushing SOE executives to pursue excess compensation. Third, the principal-agent problem in SOEs is prominent. SOE executives have gradually accumulated managerial power during the market-oriented reform, and without effective supervision, they have been empowered to pursue excess executive pay (Fang, 2009). In conclusion, the unique features and inherent complexity of the executive compensation mechanisms in SOEs have provided a fertile ground for SOE executives to pursue excessive compensation.
In order to resolve the issues aforementioned, Chinese government has introduced various regulations on SOE executive compensation. As early as 1993, the State Council began to explore the implementation of a annual compensation system for SOE executives. In 2002, the State Council stipulated that the annual compensation of SOEs executives should not exceed 12 times of the average compensation of employees. Since 2004, in order to strengthen the management of executive compensation in SOEs, the State Council has successively issued five regulatory documentations. But with fast-moving social and economic environment, theses government initiatives turned out to be futile. In 2002, the average compensation of SOE executives was capped at 12 times that of the average compensation of employees, and it increased to 13.6 times in 2003 (Fang, 2009). In 2004, the average compensation of SOE executives were higher than that of non-SOE's, with a sharp and steady rise over 2005(Li et al., 2013. Chinese government made another attempt to limit the SOE executive compensation with the until promulgation of the CRP in 2009. However, despite of the efforts in reaffirming and refining the management of deferred payment and perquisite consumption in 2004, the CRP in 2009 did not specify the requirement for total amount of executive compensation in SOEs. In 2014, the Central Committee deliberated and approved a new round of compensation regulation policies, with more rigid requirements to regulate the behaviour and compensation of SOE executives, including a ceiling standard for seven job behaviours and a ban on four kinds of public expenditures, which strictly limit the excessive compensation of SOE executives.

Literature review and hypotheses development
How will the CRP affect the political promotion of SOE executives? As mentioned above in the institutional background, due to the government support and absence of shareholders, SOE executive has been empowered to seek excessive compensation. The CRP has reduced the compensation of SOE executives, but it has not limited their managerial power. Therefore, the incentive effect of compensation on SOE executives has decreased. SOE executives, in pursuit of their self-interest, will seek possible opportunities to compensate the disbenefits caused by the institution. Further, with the policy restricting the perquisite consumption behaviours of SOE executives, the new round of SOE reform, the corruption crackdown from the central government and the incentive effect of perquisite consumption and corruption minimised, political promotion will be seen as an effective way to avoid violating the CRP and to increase the social status of individuals, which will inevitably become the priority choice of SOE executives. We anticipate that, after the inception of the CRP, SOE executives will pay more acute attention to their political future and promotion incentive will become a substitute or supplement to the compensation incentive.
The impact of compensation regulation on political promotion of SOE executives can also be explained from the perspective of behavioural economics. Behavioural economics theory posits that individuals have both economic and social preferences. For example, one would be concerned about not only his/her own compensation, but also the compensation gap in comparison with others. Therefore, the most important principle of compensation contract design is to keep it in line with the fairness alignment between agents (Fehr et al., 2007). A recently emerging theory on the Contract as a Reference Point introduces behavioural factors into contract analysis. Moreover, this theory emphasises that the contract designed in real life is rather incomplete, as it only provides a reference point for the parties signing the contract to judge the gains and losses of their own interests (Fehr et al., 2011;Hart & Moore, 2008). Therefore, according to the basic logic of Contract as a Reference Point theory, when the compensation contract deviates from the executives' expectation due to government regulation, executives will have negative psychological perception of the erosion of their own interests, and this will drive them to seek alternative incentives such as political promotion to cushion the losses incurred in the compensation contract.
As above discussed, SOE executives in China not only seek to maximise compensation as managers, but also pursue political promotion as politicians. The optimal incentive contract of SOE executives is based on the trade-off between compensation incentive and political promotion incentive. After the CRP, the compensation of SOE executives has reduced, so as the incentive effect of the compensation, and promotion incentive will become a substitute or supplement to the compensation incentive.
In addition to the views above, compensation regulation may not encourage or inhibit the pursuit of political promotion of SOE executives. Due to the nature of political promotion, i.e. long-term, uncontrollable and space-limited, it can only be the game of minorities. In addition, in the context of SOE executive compensation being regulated by the government, executives' compensation is still subject to regulation even if they are promoted. Therefore, to reduce the negative psychological perception of the erosion of their own interests as a result of the policy, some SOE executives will give up their current administrative level or reduce their endeavours to match the total compensation received. This will not necessarily encourage or inhibit the pursuit of political promotion of the SOE executives. However, the liquidity and mobility of the labour market of SOE executives is relatively low, the exit cost of executives is high, and the performance evaluation system of SOEs can prevent them from being inactive for a long time. Therefore, we still expect that compensation regulation will incentivise the high-paid executives in SOEs to actively seek political promotion.
The active pursuit of political promotion of the SOE executives will affect their behaviour, and consequently, their promotion results. We further analyse how SOE executives get promoted after the CRP. Examining the pathways through which the SOE executives can get promoted after the CRP is of great significance to understand and regulate the behaviour of SOE executives. There are generally two pathways for SOE executive in China to get political promotion. First, the performance hypothesis suggests that SOE executives would improve corporate performance as demonstration of their competence to their superiors to seek promotion Feng & Johansson, 2017). As local governments usually need to supervise several SOEs within their jurisdiction and are constrained by time, energy, ability, etc., their performance evaluation of the SOE executives is often outcome-oriented (Zhou & Zhang, 2016). For SOEs, operating performance is seen as an important manifestation of outcome to reflect the managerial competence. Therefore, in general, SOE executives with good performance are more likely to be promoted, i.e. performance is the key to the political promotion , which is the main view purported by the performance hypothesis. Second, the relationship hypothesis posits that most SOE executives would actively bear policy burdens, prioritising government agenda, to seek promotion (Huang, 2003;Zhou, 2007). With the ongoing reform, the inadequacy of social security system, the public welfare nature of social stability and the profit-pursuing nature of private enterprises, SOEs have been urged by the government to continue to bear some policy burdens and provide corresponding social welfare to help achieve some non-economic goals (Bai et al., 2000). As a result, policy burden index is considered by government when designing the promotion incentive mechanism. SOE executives also have the motivation to cater to the political preferences of higher-level government officials in order to get promoted as early as possible (Huang, 2003;Zhou, 2007). As such, policy burden can help executives to achieve political promotion (Chen et al., 2019;Liao & Shen, 2014), which is the main view purported by the relationship hypothesis. With theoretical support well-grounded in the literature for both pathways, we speculate that after the CRP, SOE executives choose to get promoted by improving corporate performance or increasing policy burdens.
Based on the discussion above, the two hypotheses are stated as follows: H1: Compensation regulation will encourage SOE executives to pursue political promotion.
H2: Compensation regulation can encourage SOE executives to seek political promotion by improving performance or increasing policy burden.

Sample selection and data sources
The sample period spans from 2013 to 2018, and the initial sample includes all stateowned enterprises from the A-share listings in China. We take the following steps to obtain the final data for empirical analysis: (1) excluding the financial companies and the ST companies; (2) excluding samples whose property rights changed during the sample period, and excluding firms admitted to the stock market after the inception of the policy; (3) excluding the sample observations with missing data. All continuous variables are winsorised at 1% and 99% to eliminate the effect of extreme values. The nominal tax rate data are collected from WIND database. The average compensation data of employees at provincial enterprises are collected from China labour statistics yearbook, and the promotion data of SOE executives are collected and merged manually from CSMAR database, annual reports of listed companies, announcements on executive changeover and other company documents, and then cross-checked through the company's official website, Baidu search and other related websites. Other financial data are collected from CSMAR database, and the data processing is completed by Stata14.0.

Research design
In pursuit of our investigations on the changes of SOE executives' pursuit of political promotion after the CRP, we construct the following model: The dependent variable, promotion, is a dummy variable representing whether SOE chairman or general manager get political promotion or not. The reason why we include both SOE chairman and general manager is that Yu and Zhang (2016) argue that the role of SOE chairman and general manager is very important for state-owned enterprises, and both are important evaluation objects of SASAC. promotion equals 1 if they get promotion and 0, otherwise. We measure the changes of political ranks of SOE executives with the following two steps: first, we divided the sample into seven categories according to the reasons for the chairman's resignation and his post-resignation designation, including lower-level SOEs, same-level SOEs, higher-level SOEs, government departments, private enterprises or other profit-making enterprises, unknown samples, and samples of chairman who quit their jobs due to health reasons and illegal crimes. For example, higherlevel SOEs means the parent company or companies at a higher level, such as municipal level SOEs, provincial level SOEs and central level SOEs. Second, we define whether the executives have been promoted politically based on the above classification. Following Yang et al. (2013) and Zhang et al. (2015, promotion can be defined as (1) transfer to a government department; (2) transfer to a higher-level SOEs; (3) transfer to a company of the same level and get a promotion. We use a dummy variable to measure the independent variable of compensation regulation (Post). According to the time of implementation of the CRP, the year before 2015 is considered as before the regulation year (Post = 0); 2015 and subsequent years are considered as after the regulation year (Post = 1). In order to measure the net effect of the CRP and eliminate the influence of other factors, we refer to Campello and Larrain (2016) and Qian & Fang (2017) and construct a control group and a treatment group in order to examine the differential impact of CRP on different SOEs. Based on this, we use DID method to test the policy effect. Specific steps are taken as follows: the CRP issued in 2015 specifically limits three components of the executive compensation at central enterprises (other state-owned enterprises refer to the implementation). First, the basic annual compensation should not exceed twice the average compensation of the provincial employees in the previous year. Second, the performance compensation should not exceed twice the basic compensation. Third, term-related incentive should not exceed 30% of the total compensation. Accordingly, the upper limit of the total compensation of SOEs executives is about 7-8 times of the average pay of the employees in provincial enterprises in the previous year (Yang et al., 2018). We divide the sample into treatment group and control group according to the compensation gap ratio calculated as the average compensation of SOE chairman and general manager over the average compensation of employees in provincial enterprises in the previous year before the CRP. If the gap ratio is equal or greater than 8, it is classified as the treatment group; if the gap ratio equals to or is less than 7, the control group. In order to reduce the noise within the treatment group and the control group, we exclude sample firms with the compensation gap ratio between 7-8 times.
The model also controls other factors that affect executive promotion, including the age of the company (Age), the size of the company (Size), the ratio of liabilities (Lev), the growth rate of the enterprise revenue (Growth), the ratio of tangible assets (Tang), the ratio of intangible assets (Intang), the ratio of inventory assets (Inv), the level of SOEs (Level), the degree of equity balance (EB), and the proportion of independent directors (Indd). At the same time, the model also controls for regional fixed effect, industry fixed effect and year fixed effect, and ε is the residual term. 4 Table 1 presents the descriptive statistics of the sample. The mean value of SOE executive promotion (Promotion) is 0.076, indicating that 7.6% of the executives of SOEs got a political promotion. The distribution of other major variables is basically consistent with the existing research.

Preliminary analysis -Did the CRP issued in 2015 work?
This paper focuses on the impacts of CRP on the compensation of SOE executives, and particularly, the promotion motivation of executives. In this part, we examine whether the CRP has affected the compensation of SOE executives. The CRP issued in 2015 specifically restricts the compensation gap between executives and employees. We use the ratio between the average compensation of SOE chairman and general manager and the average compensation of employees in provincial enterprises in the previous year to measure the compensation gap (Gap).
Firstly, we use the method of DID to test the relationship between compensation regulation and SOE executive compensation, and our results show that, after the CRP, the compensation of high-paid executive in SOEs significantly decreased, which suggests that the policy issued in 2015 has been effective. Further, we divide the whole sample into treatment group and control group to test the impact of CRP on executive compensation in different samples. The results show that the influence of CRP on SOE executive compensation in the treatment group and suggesting that the CRP has significantly reduced the executive compensation in the treatment group, and the CRP has no significant influence on the executive compensation in the control group. According to the results, we conclude that the impact of CRP on SOE executive compensation is mainly observed in the treatment group, and this also confirms that the grouping mechanism adopted is reasonable and in line with our expectation. 5 Thus far, we show that after the CRP, the incentive channels of executive compensation have been encumbered. We then move on to explore whether the political promotion of SOE executives has become an alternative incentive mechanism after the inception of this policy. Table 2 reports the impact of CRP on SOE executives' political promotion. In Table 2 column 1, the relationship is tested by DID method. The probability of political promotion of high-paid SOE executives has significantly increased after the CRP. In addition, we divide the whole sample into treatment group and control group, and then test the influence of CRP on SOE executives' political promotion in different groups respectively. The test results are shown in Table 2 columns 2 and 3. Column 2 shows this relationship in the treatment group, where the probability of political promotion of SOE executives has significantly increased after the CRP. Column 3 shows this relationship in the control group, where there is no significant change. The results in column 2 and 3 together indicate that the influence of the CRP on the political promotion of SOE executives is mainly observed in the treatment group, which supports our conjecture and excludes other possible explanations.

Basic regression analysis: compensation regulation and political promotion
Results from the basic regression analysis suggest that after the CRP, the political promotion of high-paid executives has increases significantly, and the increase of political promotion of SOE executives only occurs in the treatment group. Based on this, we conclude that after the CRP, the incentive channels of executive compensation have been encumbered, and political promotion has become an alternative incentive mechanism of compensation incentive.

Parallel trend test
In order to exclude the influence of other events before the inception of the CRP, we follow Qian and Fang (2021) and examine the changes of political promotion of SOE executives in the treatment group in comparison with those of the control group using the sample before the CRP. Specifically, we construct a time dummy variable Before (−1), which is equal to 1 in 2014 and 0 in other years, and add an interaction term between the dummy variable and Treat into the empirical model. The result shows that the coefficient of the interaction term is not significant, which supports the parallel trend hypothesis, suggesting that the conclusions of this paper are not driven by other events before the introduction of the CRP.

Dynamic effect analysis
To test the dynamic effect of CRP on political promotion of SOE executives, we construct After (0), After (1), After (2) and After (3), which are equal to 1 in 2015, 2016, 2017 and 2018 respectively, and 0 in other years. We also add interaction terms between these dummy variables and Treat into the model. The result shows that all interaction items are significant, which indicates that the political promotion behaviour of SOE executives in the treatment group was affected by the CRP in the year when it was implemented and in the following years. This again renders further support to the key findings of this paper.

Propensity score matching (PSM)
Propensity score matching (PSM) is used to reduce the characteristic variations between the treatment group and the control group. We select a series of variables for the purpose of propensity score matching from the perspectives of debt level, enterprise growth, company size, tangible assets, intangible assets, net inventory, return on assets and so on. The results are consistent with the basic regression results.

Alternative measure of the key variables
As aforementioned, we divide the sample SOEs into treatment group and control group using executive compensation gap ratio. However, the difference between the compensation of SOE executives in different groups will lead to the difference between the political promotion of SOE executives in the treatment group and the control group before the CRP due to unobservable factors, thus reduces the effectiveness of the DID model. To address this concern, we reselect the treatment group and the control group to minimise the impact caused by the compensation differences. Specifically, we limit the executive compensation gap to 8-9 times in the treatment group and 6-7 times in the control group. The results are consistent with the basic regression results.

Random grouping test
To eliminate the interference of other external factors and strengthen the causal relationship between the CRP and SOE executive political promotion, we randomly divide the sample into treatment group and control group and repeat the simulation experiment for many times. If the results are not significant, it indicates that the placebo test is successful, that is, other external factors have little influence on our key results. To this end, the randomised experiment was repeated 500 times. The result shows that the estimated coefficient is distributed near zero and display characteristics of a normal distribution, which means that the placebo test has been passed. In other words, the exogenous event of CRP is the main influencing factor for the increase of political promotion of SOE executives.

Fictitious policy time
If the relationship between the CRP and the political promotion of SOE executives is driven solely by the inception of this policy, the results from DID analysis should be insignificant when the event year is changed artificially. For this purpose, we push forward the time of the total sample interval and the CRP by 3 years, and then rerun the DID test. We find that the significant influence of the fictitious CRP on SOE executive political promotion has disappeared, suggesting that the increase in political promotion of SOE executives documented by the previous DID test is indeed driven by the implementation of the CRP.

Eliminate the influence of other factors
The political promotion of SOE executives may also be affected by the change of government, that is, the increase in political promotion of executives is not caused by salary control, but caused by a large number of personnel changes brought about by the change of government. 2018 is the year of the general election, and the local two sessions will complete the centralised election of the provincial people's congress, the government and the Chinese People's Political Consultative Conference. In order to exclude the influence of government change on the conclusions of the result, we remove the year of change.The results are consistent with the basic regression results.

Mechanisms
The analysis above shows that the pursuit of political promotion of high-paid SOE executives has increased after the introduction of the CRP. We further analyse how SOE executives achieve the political promotion when their compensation is regulated. As previously discussed, there are usually two pathways for SOE executives to get promotion, either via improving corporate performance Feng & Johansson, 2017) or by bearing policy burdens (Huang, 2003;Zhou, 2007). The second objective of this paper is to examine through which pathways SOE executives can get promoted after the CRP, and this is of great significance to shed light on the behaviour of SOE executives and offer insights to future regulations.

Relationship hypothesis: executives get promotion mainly by catering to the government needs
Corporate tax burden and policy burden are two important channels for SOE executives to cater to the government needs (Bradshaw et al., 2019;Dai et al., 2017;Chen et al., 2019). After the reform of tax sharing system, local governments started to pursue fiscal surplus (Feng & Shen, 2015), and the tax planning has evolved into a compulsory political task for the local governments (Zheng, 2005), with hard binding force (Gao, 2006). For local government officials, the contribution of tax revenue to political promotion is greater than that of the economic growth (Bo, 1996). In addition to tax revenue, reducing unemployment and maintaining social stability are also important objectives of the government, which are largely shifted to SOEs as policy burdens in their jurisdictions. Therefore, we expect that after the CRP, SOE executives will be more inclined to cater to the government needs by shouldering more corporate tax burden and policy burden so as to get promoted. We use corporate tax burden and policy burden to measure the ingratiatory behaviour of SOE executives. Following Wu (2009) andBradshaw et al. (2019), corporate tax is calculated as Tax = Cash Income Taxes Paid/Pre-Tax Income. We follow Lin et al. (2004), Bai and Lian (2014) to measure policy burden.
We first test the impact of the CRP on the ingratiatory behaviour of SOE executives. The results are shown in Table 3. The dependent variables in columns 1 and 3 are enterprise tax burden and enterprise policy burden respectively. Our results show that the CRP has significantly increased the ingratiatory behaviour of SOE executives in the high-paid group. Then, we examine how CRP affects the relationship between corporate tax burden, policy burden and political promotion. The results are shown in columns 2 and 4. Column 2 shows the results of corporate tax burden as an influence channel, the coefficient of which is significantly positive, indicating that the higher the corporate tax burden is, the more likely it is for executives to achieve political promotion. Furthermore, the coefficient of Post×Treat×Tax is also significantly positive, indicating that the CRP has strengthened the pathway to political promotion of SOE executives through shouldering more tax burden. Column 4 shows the results of policy burden as an influence channel, and the CRP has also strengthened the pathway to political promotion of SOE executives through taking on more policy burden. The results above suggest that after the CRP, the executives in the high-paid group have engaged in more ingratiatory behaviours, and more executives have resorted to seek political promotion. Also ingratiatory behaviour is an important channel for SOE executives to obtain political promotion when their compensation is regulated.

Performance hypothesis: executives get promotion mainly by improving corporate performance
We use the same method above to test performance hypothesis, we anticipate that SOE executives in high-paid group will seek political promotion by actively improving corporate performance after the CRP.
We use the same method above to test this. Table 4 column 1 shows the change of SOE performance after the CRP, where CRP has significantly reduced the performance of SOEs. Our results do not render support to the performance hypothesis. Furthermore, the coefficient of Post×Treat×Roa is not significant, indicating that improving corporate performance is not an important channel for SOE executives to obtain political promotion when their compensation is regulated.
Based on the above, we conclude that ingratiatory behaviour is the main channel for SOE executives to seek promotion after the CRP, and this supports the relationship hypothesis in relation to the political promotion of the SOE executives.

Further evidence: compensation regulation and political promotion
Our analysis above shows that after the CRP, SOE executives will pursue political promotion by catering to the government needs. In this section, we use three types of grouping variables to verify the influence of CRP on political promotion of the SOE executives under different circumstances.

Executive age
Although the central government does not specify a particular age at which officials can be promoted, the Leading Cadres Retirement Provision sets out the age limit at which officials of different levels can be promoted. In practice, there exists a capping age threshold for political promotion. In addition, most SOEs actively promote the principle of recruiting cadres at a younger age. Young SOE executives actively capitalise on their age advantage through their career progression, because as they get older, the proportion of promotion incentives in the overall incentive structure will decrease (Jiang et al., 2015). Therefore, we expect that if compensation regulation affects the political promotion motivation of SOE executives, such effect should be more prominent amongst the younger managers. We sort the sample according to the age of SOE executives, and executives with above-median age are classified as the high-age group; otherwise, the low-age group. The results are shown in Table 5. The coefficient of Post×Treat×Eage is significantly negative, indicating that the CRP has a greater impact on the political promotion of SOE executives in the low-age group, which provides support to our conjecture above. In addition, we separate the sample into different age groups and our results are robust that CRP significantly improves the political promotion of SOE executives only in low-age group.

Corporate level
In China, SOEs can be divided into central level SOEs and local level SOEs. There are significant differences between them in relation to their operating objectives, degree of government intervention, restrictive incentive mechanism and regulatory independence (Hao & Liu, 2009). Under the current incentive model of promotion tournament for government officials in China, local government, as the ultimate controller of the local level SOEs, has the right to appoint managers, and local government officials have strong incentives to intervene in the local level SOEs in order to achieve social and political goals such as expanding fiscal revenue, increasing GDP, improving social welfare and maintaining social stability (Lin et al., 2004;Zhou, 2007). SOE executives would actively cater to the political preferences of higher-level government officials to secure a swift promotion (Huang, 2003;Zhou, 2007). And To empirically test this, Post×Treat×Level is added into the model. When the enterprise is a central level SOEs, Level is equal to 0; when the enterprise is a local level SOEs, Level is equal to 1. The results are shown in Table 6. The coefficient of Post×Treat×Level is significantly positive, indicating that the CRP has a greater impact on the political promotion of local level SOE executives. We also separate SOEs at different levels as further analysis, where our results remain robust that CRP significantly improves the political promotion of SOE executives only in local level SOEs.

Regional marketisation degree
We further explore how the degree of regional marketisation affects the relationship between compensation regulation and political promotion. We expect that the lower the degree of marketisation of the region, the stronger the motivation and ability of the government to intervene in the SOEs, and the more likely that SOE executives will cater to the government needs to seek promotion opportunities after the CRP.
To test this hypothesis, we group the regions according to the upper quartile of the regional marketisation index (Mar). If the regional marketisation index is higher than the upper quartile of the national marketisation index, it is considered as region with higher degree of marketisation; otherwise, a lower degree of marketisation. We first add the Post×Treat×Mar variable to the model, and the results are shown in Table 7. The coefficient of Post×Treat×Mar is significantly negative, meaning that the CRP has a more significant impact on the political promotion of SOE executives in regions with a lower marketisation degree. We also separate regions with different degrees of marketisation and our results remain robust.

Compensation regulation, market-oriented reform of SOEs and political promotion
Thus far, we show that regional market processes have weakened the relationship between compensation regulation and political promotion. In this section, we analyse how the introduction of market factors in enterprises will affect this relationship. Understand the economic consequences of the superposition of the past reform policies of SOEs will provide insights for future regulation promoting SOE reform in China.
We use the scenario of the mixed ownership reform of SOEs to contextualise such an investigation. The mixed ownership reform of state-owned enterprises has reduced government intervention in SOEs and restrained the agency problem of SOE executives . We expect that the mixed ownership reform of SOEs will weaken the relationship between compensation regulation and political promotion. To verify the above inference, we add the Post×Treat×Mix variable to the model, and the results are shown in Table 8. The coefficient of Post×Treat×Mix is significantly negative, suggesting that the CRP has a greater impact on the political promotion of SOEs with a lower mixed ownership reform of SOEs degree. In addition, we also separated SOEs with different degrees of mixed ownership reform of SOEs to verify these results, and the results are Unchanged.  Table 7. Compensation regulation, regional marketisation degree and political promotion.

Political Promotion
Full sample Higher Regional marketisation degree

Compensation regulation, other benefit acquisition channels and political promotion
In the analysis above, we have examined the relationship between compensation incentive and political promotion incentive, and showed that the reduction of compensation will encourage SOE executive to pursue political promotion. In this section, we further introduce variables like equity incentive (Mshare) and perquisite consumption (Peak) to explore how other benefit acquisition channels affect the relationship between compensation regulation and political promotion of SOE executives. The optimal incentive contract for SOE executives is not only relating to the trade-off between compensation incentive and political promotion incentive, but also depending on how many resources the executives can obtain from other benefit channels. In another word, whether lower compensation prompts SOE executive to pursue political promotion also depends on the other benefit acquisition channels possessed by these executives. If the benefit acquired from other channels could make up for the compensation losses incurred, SOE executives would less likely to pursue political promotion after their compensation was reduced under the regulation. Therefore, we propose that a higher number of shareholdings and perquisite consumption would mitigate the relationship between compensation regulation and political promotion of SOE executives.  Table 9. Compensation regulation, other benefit acquisition channels and political promotion.
(1) To verify the above inference, we add Post×Treat×Mshare and Post×Treat×Perk to the model, and the results are shown in Table 9. The coefficients of Post×Treat×Mix and Post×Treat×Perk are significantly negative, which has confirmed our prediction above that a higher number of shareholdings and perquisite consumption would weaken the relationship between compensation regulation and political promotion of SOE executives. We also try different ways of separating SOE executives with various shareholdings and perquisite consumption to rerun the test, and our results remain robust.

Conclusion
Using a sample of listed state-owned enterprises in China from 2013 to 2018, this paper focuses on the impacts of CRP, as an exogenous shock event, on the change of incentive mode for SOE executives. Unlike executives in non-SOEs and other countries, SOEs executives in China pursue not only compensation incentive as managers, but also political promotion as politicians. We find that after the inception of the CRP, the compensation incentive channel for highly paid executives has been encumbered and has been replaced by political promotion as an alternative incentive mechanism. Further analysis shows that ingratiatory behaviour has become the main channel for SOE executives to seek promotion after the CRP. Our results provide support for the relationship hypothesis relating to the political promotion of SOE executives. In addition, we introduce the grouping variables of executive age, corporate level and regional marketisation degree to further explore the influence of compensation regulation on executive promotion under different circumstances. Our results show that the effects above are more pronounced within younger SOE executives, lower enterprise level and lower marketisation degree. Finally, we find that the existence of alternative benefit acquisition channels such as equity incentive, perquisite consumption and the degree of mixed ownership reform of SOE has weakened the relationship between compensation regulation and political promotion of SOE executives. In conclusion, this paper reveals the important role of political promotion as an executive incentive mode and provides insightful support for further SOE reform in China.
Our results suggest that the regulation of SOE executive compensation has brought changes to the executive incentive mechanism, consequently, affecting the behaviour of executives and corporate performance. Such finding has not only shed further light on the economic consequences of the CRP at the micro level, but also provided insights to significance and future reference for the reform of SOEs in China going forward. Taken collectively, further adjustment of the promotion mechanism and assessment mechanism of SOE executives forms an integral part in the course of promoting long-term and sustainable development of SOEs. What is more important is to change the government's appointment system for the SOE executives and allow them to return to the single status as managers. Fundamentally, to improve the efficiency of SOEs, senior executives should be selected following the market mechanism according to their operational capabilities, with the influence of the government on SOE executives minimised.

Disclosure statement
No potential conflict of interest was reported by the author(s).