Is the corporate social responsibility countercyclical? – evidence from Chinese listed companies

ABSTRACT Whether the real performance of corporate social responsibility is related to social progress and development in the critical period of China’s economic transformation. It is crucially significant for improving social welfare and promoting high-quality economic development to clarify the relationship between the business cycle and corporate social responsibility. Choosing A-share listed companies in China’s capital market between 2011 and 2018 as a sample, we find corporate has a better social performance during business contraction, and social responsibility presents a counter-cyclical support effect, which is more obvious in non-state corporate. Public popularity, growth level and financial distress can significantly promote this counter-cyclical support effect. Further research finds that social responsibility in economic contraction can bring government subsidy, financing convenience and reputation return for corporate in the next year. Our study enriches the literature on business cycle and corporate social responsibility and promotes high-quality economic development.


Introduction
China's economy has shown a steady growth trend and the pace of structural adjustment has been stable since the supply-side reform.However, the negative impact of short-term self-interest behaviour has gradually expanded with the development of many enterprises, problems such as environmental pollution, infringement of the interests of small and medium shareholders, cheating consumers, and food safety often occur.For example, the 2015 Hubei Changyang Chemical Pollution Case, the 2018 Changchun Changsheng 19th Central Committee of the Communist Party of China clearly stated: We need to accelerate the reform of the ecological civilisation system and build a beautiful China.We need to raise pollution discharge standards, strengthen polluters' responsibilities, and improve systems such as environmental credit evaluation, mandatory disclosure of information, and severe punishment.Food safety national standard Twelfth Five-Year Plan also clarifies the objectives, tasks and priorities of food safety construction standards.This system and realistic background reflect the urgency and importance of corporate social responsibility, which is closely related to high-quality economic development.It is of theoretical significance and practical value to explore the relationship between the economic cycle and corporate social responsibility from the perspective of high-quality development in the critical period of China's economic transformation.
The behaviour decision of the enterprise is directly decided by the internal management, but it is also affected by the external macro environment.Economic cyclical fluctuations may affect managers' decision-making and external business environment, and then affect corporate behaviour strategy.Scholars have studied the impact of the economic cycle on the micro-behaviour of enterprises.For example, Lian et al. (2020) study the peer effect of the capital structure under the economic cycle.D. Chen et al. (2016) analyse the impact of the economic cycle on the tax avoidance of state-owned enterprises.Begenau and Salomao (2019) examine the changes in debt financing and equity financing of companies of different sizes under the economic cycle.However, it is regrettable that the existing literature mainly studies the economic cycle from the aspects of corporate financing, investment level, capital structure adjustment and tax avoidance.Social responsibility, the micro-behaviours of enterprises, is rarely discussed.So it becomes the starting point of this paper.
We analyse that corporate social responsibility response to the economic contraction cycle is a defensive effect or support effect.Under the economic contraction cycle, corporate resources are reduced and corporate strategies are conservative.To retain funds to enhance financial flexibility, the corporate is likely to reduce social responsibility investment.Corporate social responsibility response to the economic contraction cycle is the defensive effect.Moreover, Corporate social responsibility response to the economic contraction cycle may be a support effect.Under the economic contraction cycle, companies may actively take on social responsibility to obtain certain government subsidies, financing facilities and reputational rewards considering that social responsibility is lowcost, low-risk and rewarding.The above logic remains to be further empirical analysis below.Therefore, we intend to explore the relationship between the economic cycle and corporate social responsibility.
We take A-share listed companies in China's capital market from 2011 to 2018 as samples, and empirically examine the relationship between economic cycle and corporate social responsibility.We further analyse the situational effects of factors such as the nature of property rights, public popularity, growth level and financial distress as well as the economic consequences of enterprises changing with economic cycle.The results show that corporate has a better social performance during business contraction, and social responsibility presents a counter-cyclical support effect, which is more obvious in nonstate corporate.Public popularity, growth level and financial distress can significantly promote this counter-cyclical support effect.Further research finds that social responsibility in economic contraction can bring government subsidy, financing convenience and reputation return for corporate in the next year.
Compared with the previous literature, the possible contributions are mainly reflected in the following three aspects.First, our study not only enriches the literature in the field of economic cycle, but also has a certain reference role in promoting social welfare and high-quality economic development.The existing research mainly examines the economic cycle from the perspective of micro-enterprise behaviour such as financing, investment, capital structure adjustment and tax avoidance (Begenau & Salomao, 2019;D. Chen et al., 2016;Lian et al., 2020).We discuss economic development from the perspective of high-quality development based on the perspective of corporate social responsibility.Secondly, our study deepens and expands the literature on the influencing factors of corporate social responsibility.The existing literature mainly examines the impact of internal and external governance mechanisms on corporate social responsibility (Eberhardt-Toth, 2017;Ghoul et al., 2019).We analyse the fulfilment of corporate social responsibility from the perspective of the economic cycle based on the current state of macroeconomic development.Third, our study enriches the understanding of issues related to the macroeconomic cycle, and also provides some reference for the government and enterprises to supervise and control corporate social responsibility in different situations.Because we examine the situational effect of the nature of property rights, public popularity, growth level, and financial distress.

Literature review
As an important manifestation of social progress and high-quality economic development, corporate social responsibility has gradually attracted the attention of academic and practical circles since the 20th century.The existing literature shows that corporate social responsibility has both value creation and tool effects.On the one hand, corporate social responsibility can improve corporate reputation (Turban & Greening, 1997), enhance the brand effect, gain consumer loyalty (Daub et al., 2005), gain government subsidies (Du & Chen, 2016) and ease financing constraints (Gao et al., 2017).Thus it can improve corporate performance (Porter & Kramer, 2006) and produce a significant valuecreation effect.On the other hand, the manager may also be to divert the attention of external or internal personnel and cover up their bad self-interested behaviour by corporate social responsibility.Corporate social responsibility has a certain instrumental effect (Petrovits, 2006).Therefore, more scholars have gradually turned their attention to the influencing factors of social responsibility based on the contribution and value of social responsibility.Domestic and foreign scholars mainly analyse how to improve the performance of corporate social responsibility to obtain more value returns from the perspectives of enterprise characteristics, consumers and external environment.
The topic related to our study is the external environment.Scholars find that social trust, media freedom, national culture, and voluntary disclosure policies can affect corporate social responsibility.For example, Yang et al. (2021) find that social trust can significantly promote corporate social responsibility.The market environment can significantly improve the effectiveness of social trust in promoting corporate social responsibility.The equity incentive executives can significantly inhibit the improvement of social trust in corporate social responsibility.Ghoul et al. (2019) find that media freedom plays a certain role in information dissemination.Taking 42 countries as an example, the study finds that national media freedom can significantly promote corporate social responsibility.The above effect is more obvious in enterprises with high governance levels and large scales.Peng et al. (2014) find a significant relationship between national culture and corporate social responsibility.Specifically, individualism and uncertainty avoidance can significantly improve corporate social responsibility.Power distance and masculinity have a certain inhibitory effect on corporate social responsibility.Lone et al. (2016) use the voluntary guidelines for corporate social responsibility as an entry point, and find that after the introduction of the principle, social responsibility disclosure and performance are significantly improved.The proportion of independent directors, female directors and board size can promote the above effect.In addition, social corruption and moral decline (Zhengm & Lu, 2018), national audit (Pan et al., 2020), and business ethics (Manasakis, 2018) also can promote corporate social responsibility performance.
It can be seen that few literature studies on how to change the motivation and behaviour of corporate social responsibility in the case of cyclical fluctuations in the macro-economy.Although there are many kinds of research on the influencing factors of corporate social responsibility.Therefore, we intend to explore the relationship between the economic cycle and corporate social responsibility from the perspective of high-quality development.It enriches the literature on the economic cycle and corporate social responsibility and helps the government to reasonably grasp the macroeconomic laws, supervise corporate social responsibility, and promote high-quality social development.

Theoretical analysis and research hypothesis
The world economy has suffered more or less negative shocks since the outbreak of the financial subprime mortgage crisis in 2008.How the macro economy affects the micro behaviour of enterprises has gradually attracted the attention of scholars all around the world.The economic cycle is a phenomenon in which economic expansion and economic contraction alternate.Compared with the economic expansion cycle, the market development is depressed and the corporate has higher financing pressure (Begenau & Salomao, 2019;Luo & Gong, 2014), and higher tax avoidance (D.Chen et al., 2016) during the economic contraction cycle.Meanwhile, the corporate is more inclined to adopt stable business strategies, such as shrinking the market and making dynamic adjustments to the company's capital structure (Qian & Hu, 2015).The cyclical fluctuations of the macro economy could affect the company's dynamic adjustment and managers' decision-making judgement, thus affecting the micro-enterprise behaviour.As one of the micro-enterprise behaviour, corporate social responsibility is not only the embodiment of the role of corporate social citizenship but also a capital investment.It may be affected by external economic fluctuations.In the economic contraction cycle, corporate social responsibility may present two different changes.Corporate social responsibility response to the economic contraction cycle may be a defensive effect or support effect.
Corporate social responsibility response to the economic contraction cycle may be a defensive effect.Corporate social responsibility is lower in the economic contraction cycle.In the economic contraction cycle, access to corporate resources is reduced and their strategies are more conservative.Therefore, the corporate can reduce social responsibility investment to retain funds, enhance financial flexibility and improve the ability to defend against operational uncertainty.Specifically, Corporate can reduce social responsibility investment in the economic contraction cycle from perspectives of resources and managers' psychological characteristics.On the one hand, the market environment is depressed in the economic contraction cycle.It could lead to business contraction, lower operating performance, and earnings (Qian & Hu, 2015).Meanwhile, the national monetary policy shrank in the economic contraction cycle.Banks and other credit institutions are also more stringent in auditing.It can increase the cost of external financing for enterprises and reduce corporate financing opportunities (Luo & Gong, 2014).The above two reasons together cause the decrease in enterprise resources and cash flow.Social responsibility, a resource consumption activity, requires the support of certain resources.Enterprises lacking funds or resources have a weak motivation to undertake social responsibility (J.Chen & Zheng, 2020).Therefore, the reduction of resources in the economic contraction cycle could lead to a decline in corporate social responsibility (J.Chen & Zheng, 2020).On the other hand, in the economic contraction cycle, the product market is depressed and the external uncertainty is higher.It could enhance the difficulty for managers to predict the future market development trend (R. Zhang et al., 2011) and reduce confidence to predict the future business direction (Zhou et al., 2021).Therefore, managers could adopt conservative strategies and reduce social responsibility to retain funds and respond to business uncertainty.
Based on the above two aspects, enterprises can reduce the social responsibility investment in the economic contraction cycle.The purpose is to retain funds and cash flow to enhance their financial flexibility and respond to business uncertainty.Corporate social responsibility response to the economic contraction cycle is a defensive effect.We propose Hypothesis H1a: H1a: Limited by other conditions, corporate social responsibility response to the economic contraction cycle is a defensive effect.Enterprise has worse social responsibility performance in the economic contraction cycle.
Corporate social responsibility response to the economic contraction cycle also may be a support effect.Enterprises could actively enhance social responsibility investment and share the social development pressure in the economic contraction to ensure the highquality and stable development of society.
In the economic contraction cycle, the product market is depressed and the external uncertainty is higher.The business performance is also worse and the company faces financing constraints or financial difficulty (Qian & Hu, 2015;Wu et al., 2015).Managers are more inclined to adopt the conservative strategy and investment tendency due to the high external risk.How to solve their crisis under low risk, ease financing constraints and financial pressure to ensure sustainable development is an important issue for enterprises to consider.Existing literature shows that fixed asset investment, innovation investment and other risks are high.So enterprises may adopt defensive strategies to reduce fixed asset investment under higher market uncertainty (Huang et al., 2021).In contrast, the social responsibility investment is lower in cost and risk, and has a certain sustainable return (Porter & Kramer, 2006).So it may become the better choice in the economic contraction cycle.
Compared with fixed asset investment and innovation investment, social responsibility investment is lower in cost and risk, and has a certain sustainable return.We believe that enterprises may enhance social responsibility investment based on value creation motivation in the economic contraction cycle.Corporate social responsibility response to the economic contraction cycle is support effect.
Specifically, social responsibility can bring three benefits to enterprises, such as government subsidies (Du & Chen, 2016), financing facilities (Gao et al., 2017) and reputation return (Daub et al., 2005;Turban & Greening, 1997).Therefore, in the economic contraction cycle, companies may enhance social responsibility investment based on value creation motives.First, in the economic contraction cycle, companies may enhance social responsibility investment to obtain government subsidies.In the economic contraction cycle, companies face worse performance and financial pressure.To obtain government subsidies to ease their financial pressure, companies need to meet government needs to a certain extent (Li et al., 2016).Therefore, enterprises may enhance social responsibility investment to help the government revive the economy.They actively play the role of social citizens and get government favours to gain future government subsidies.Secondly, in the economic contraction cycle, companies may enhance social responsibility investment to obtain financing facilities.In the economic contraction cycle, the market credit policy and external financing environment are worse.Enterprises may face more serious financing constraints (Luo & Gong, 2014).Corporate social responsibility investment could bring financing convenience, and ease the pressure on corporate finance (Gao et al., 2017).Therefore, enterprises may enhance social responsibility investment to obtain financing facilities.Lastly, in the economic contraction cycle, companies may enhance social responsibility investment to gain consumer recognition and build brand reputation.Corporate performance deteriorates in the economic contraction cycle.How to compete in the consumer market and improve performance is also the focus of attention at this time.The existing literature shows that reputation reflects the excellent ability and good products of the corporate.It is also competitiveness that can improve performance and promote the development of enterprises (Roberts & Dowling, 2002).Therefore, enterprises may enhance social responsibility investment to win the favour of consumers and obtain a reputational return.
Based on the above analysis, we believe that enterprises may enhance social responsibility investment in the economic contraction cycle based on the social responsibility value creation motivation of government, financial institutions and consumers.Corporate social responsibility response to the economic contraction cycle is support effect.We propose Hypothesis H1b: H1b: Limited by other conditions, corporate social responsibility response to the economic contraction cycle is support effect.Enterprise has better social responsibility performance in the economic contraction cycle.

Model setting
To test the relationship between the economic contraction cycle and corporate social responsibility, we construct Model (1): In Model (1), GAP i,t-1 represents the economic cycle variable of period t-1.Referring to Fang and Chen (2019), we use the macro-economic HP filtering method to measure it and generate variables GAP i,t-1 and GAPD i,t-1 .Specifically, we use the HP filtering method to eliminate the long-term trend of China's GDP growth rate from 2010 to 2017.We obtain the real GDP growth rate gap GAP to measure the economic cycle from 2011 to 2018.Then we gain the variable GAP i,t-1 and GAPD i,t-1 .To better interpret the regression results, the variable GAP i,t-1 is equal to the negative number of the actual growth rate gap.GAPD i,t-1 represents whether the economy is in a contraction cycle.If GAP i,t-1 is greater than 0, GAPD i,t-1 is assigned 1; if GAP i,t-1 is less than 0, GAPD i,t-1 is assigned 0. In the 2010 to 2017 cycle, the actual year of economic contraction is 2011 to 2015.2010, 2016 and 2017 are years of economic expansion.It is basically consistent with the years' summary of other scholars (M.Chen & Zhang, 2016;Ding & Yan, 2019), indicating that the variables are rational.
CSR i,t represents the social responsibility of enterprise in the t period.Referring to the research (X.Chen, 2021), we use the corporate social responsibility index and rating to measure from Hexun Information Technology.Then we gain the variables of CSR1 i,t and CSR2 i,t .CSR1 i,t is equal to corporate social responsibility index divided by 100.CSR2 i,t is set to 3, 2 and 1 from high to low according to rating A, B and C. The 0 is defined as other lower ratings.As a third-party institution, Hexun Information Technology is independent and professional.The corporate social responsibility index and rating have high objectivity.The social responsibility data of listed companies are relatively complete.
Specifically, these control variables include: SIZE is the natural logarithm of total assets at the end of the year; LEV is the total liabilities divided by total assets at the end of the year; TOP1 is the percentage of shares held by the first shareholders divided by total shares; ROA is the rate of business revenue growth; GROW is the rate of sales growth; INDEP is the ratio of number of independent directors to the number of board directors; BOARD is the natural logarithm of board directors numbers; SALA is the natural logarithm of total amounts of the top three executives' salary; HOLD is equal to the ratio of shares held by senior executives to total shares; AGE is equal to the average age of senior executives; GENDER is equal to proportion of women in the senior management team; ENERGY is defined as 1, otherwise 0 if the enterprise is a listed company involving new energy concepts such as wind energy, solar energy, nuclear energy, hydrogen energy, geothermal energy and biomass energy based on the new energy concept section of the Tonghuashun Financial website.Finally, we include industry and area fixed effects in the model to control for the unobserved factors that might drive the results.Considering that the economic cycle variables are time series data, controlling directly year fixed effects may produce multicollinearity.So we control macro variables to replace year fixed effects (C.Zhang & Liu, 2018).We further control the variables of monetary policy(HBZC i,t-1 ), capital market development(DEVE i,t-1 ) and natural disaster situation(DISATER, t-1 ).There are equal to the growth rate of money supply (M1), the annual turnover rate of stocks and the natural logarithm of annual disaster losses.

Sample selection
We take A-share listed companies in China's capital market from 2011 to 2018 as research samples.The actual sample interval is 2010 to 2017 due to the lag of the explanatory variables of the model.The reasons for selecting this sampling interval are as follows.First, China's economy has been affected by the financial crisis from 2007 to 2009.Secondly, the revision of the Guidelines for the Governance of Listed Companies added a chapter on stakeholders, environmental protection and social responsibility in September 2018.Enterprises and society pay more attention to the environment and ESG, which may have a certain impact on the social responsibility score.We screen the relevant data to avoid the adverse effects of abnormal samples.First, we delete the samples of abnormal trading states such as ST and PT.Secondly, we delete the samples of the financial and insurance industry.Finally, we delete the samples with missing data.After screening, we remain17951 samples.In addition, we also shrink all continuous variables at 1% and 99% quantiles.The macroeconomic data and financial news data are from the China Economic Network Economic Statistics Database.The corporate social responsibility data is from Hexun Information Technology.Other data are from the China Stock Market & Accounting Research Database(CSMAR).The analysis software is Stata 15.0.

Descriptive statistics and correlation analysis
Table 1 reports the descriptive statistical results of the main research variables.The mean and standard deviation of CSR1 i,t are equal to 0.2455 and 0.1708, respectively.The maximum value is 0.7509 and the minimum value is −0.0447.It indicates that there is a big difference in corporate social responsibility scores between different listed companies.The maximum and minimum values of GAP i,t-1 are 3.0086 and −5.1625, respectively.The standard deviations are 2.5160.It indicates that there are certain cyclical fluctuations in China's economy.The mean of TOP1 i,t-1 is 0.3522 and the maximum value is 0.7510.It indicates that the average shareholding ratio of the largest shareholder of the sample company is 35.22% and the maximum is 75.10%.This result reflects the fact that the equity concentration of Chinese listed companies is still high after the split share structure reform.The mean of INDEP i,t-1 is 0.3734.It indicates that 37.34% of the directors are independent directors in samples, which is aligned with the requirement of the China Securities Regulatory Commission that the proportion of independent directors should not be less than 1/3 of total board directors.The mean and maximum of HOLD i,t-1 are 0.0691 and 0.6208.It indicates that the average shareholding ratio of executives in listed companies is low, but the shareholding ratio of executives in several companies is large.The mean value of AGE i,t-1 is 46.7637.It indicates that the average age of executives of listed companies is close to 47 years old.In addition, the mean of GEND i,t-1 is 0.1532.It indicates that 15.32% of the listed company's executives are female.
Table 2 reports the Pearson correlation coefficient of main research variables.The Pearson correlation coefficient between GAP i,t-1 and CSR i,t is 0.0476, which is significantly positive at the 1% level.The correlation coefficient between GAPD i,t-1 and CSR i,t is 0.1271, which is significantly positive at the 1% level.It shows that there is a positive correlation between the economic contraction cycle and corporate social responsibility.The above correlation analysis results initially support the research hypothesis H1b.

Empirical results
Table 3 reports the result of multiple regression analysis of economic contraction cycle and corporate social responsibility.The explanatory variables are the GAP i,t-1 and GAPD i,t-1.
Columns (1) and ( 2) are the CSR1 i,t and CSR2 i,t , respectively.We adjust the T and Z values of all reports by Cluster method at the firm level.In Table 3, the coefficients between GAP i,t-1 and CSR1 i,t and CSR2 i,t are 0.0046 and 0.0447 respectively, which are significantly greater than 0 at the 1% level.The coefficients between GAPD i,t-1 and CSR1 i,t and CSR2 i,t are 0.0878 and 1.9683 respectively, which are significantly greater than 0. It indicates that enterprises have better social responsibility performance in the economic tightening cycle.The empirical evidence supports the research hypothesis H1b.The market operating environment is worse and the business of enterprises is shrinking in the economic contraction cycle.Performance of enterprises and financing ability are also negatively impacted.To obtain government subsidies, financing facilities and reputation returns, and alleviate the difficulties, enterprises can enhance social responsibility investment to help society recover economically.Corporate social responsibility response to the economic contraction cycle is support effect.Among the control variables, the coefficient of SIZE i,t-1 is significantly positive, indicating that larger assets help corporate to improve social responsibility.The larger the scale of the enterprise, the higher the level of social responsibility.The coefficient of ROA i,t-1 is significantly positive, indicating that the higher the corporate profitability, the better the social responsibility performance.The coefficient of GROW i,t-1 is significantly negative, indicating that the higher the level of corporate growth, the worse the social responsibility performance.The coefficient of BOARD i,t-1 is significantly positive, indicating that the larger the board, the better the social responsibility performance.The coefficient of INDEP i,t-1 is significantly positive, indicating that the higher the proportion of independent directors, the better the social responsibility performance.The coefficient of HOLD i,t-1 is significantly negative, indicating that the shareholding ratio of executives can inhibit corporate social responsibility.

Sensitivity test
We use the following series of sensitivity tests to improve the robustness of the research conclusions.

Changing measurement method of corporate social responsibility
We use corporate social responsibility index and rating from Hexun Information Technology to measure the corporate social responsibility in the previous study.To make the research results more robust, we change the measurement method of social responsibility.First, we calculate whether corporate has the charitable donation and environmental investment(CSR3 i,t and CSR4 i,t ).If the enterprise has charitable donations, CSR3 i,t is defined as 1, otherwise 0. If an enterprise has an environmental investment, CSR4 i,t is defined as 1, otherwise 0. Secondly, referring to Quan et al.(2015), we further use Rankins CSR Ratings to measure it.To eliminate the difference in magnitude, we use the Corporate Responsibility Index divided by 100 to measure the corporate social responsibility(CSR5 i,t ).Table 4 reports the empirical results.The coefficients between GAP i,t-1 (GAPD i,t-1 ) and CSR3 i,t (CSR4 i,t , CSR5 i,t ) are all significantly negative.The above results have not changed substantially compared with the above empirical analysis.It indicates that the research results are robust.

Instrumental variable method
To reduce the interference of endogenous explanatory variables on the research conclusions, we use the instrumental variable method to test the model.Specifically, referring to the economic cycle measurement method, we choose the filtered of the growth rate of per capita power consumption and the growth rate of regional highway mileage in one lagging period(DLZZ i,t-1 and GLZZ i,t-1 ) as instrumental variables.Both are continuous variables, with standard deviations of 2.2750 and 1.2652, respectively.Variable values are volatile.The reasons for choosing instrumental variables are as follows.First, electricity consumption, an important manifestation of improving people's livelihoods, can significantly promote economic growth (He et al., 2018).The growth of per capita electricity consumption could reflect the speed of economic expansion or contraction.It has a positive effect on the economic cycle variable(real GDP growth rate gap).The higher the per capita electricity consumption growth, the greater the real GDP growth rate gap.Similarly, the longer the highway mileage, the lower the transportation cost.It can promote economic growth (Tang et al., 2021;X. Zhang et al., 2018).Therefore, per capita electricity consumption and highway mileage growth can meet the correlation hypothesis of instrumental variables.Secondly, per capita electricity consumption and highway mileage can hardly directly affect corporate social responsibility in the short term.It meets the exogenous hypothesis of instrumental variables.
Table 5 and 6 report the empirical results of instrumental variable method.The coefficients between DLZZ i,t-1 (GLZZ i,t-1 ) and GAP i,t-1 (GAPD i,t-1 ) are significantly negative in Table 5.It indicates that the growth of per capita electricity consumption and the growth of highway mileage are significantly negatively correlated with the economic contraction cycle.The results conform to the correlation hypothesis of instrumental variables.The coefficients between DLZZ i,t-1 (GLZZ i,t-1 ) and GAP i,t-1 (GAPD i,t-1 ) in Table 6 have not changed substantially compared with the above empirical analysis.We test the rationality of instrumental variables.In Table 6, the partial F statistics are 601.63 and 2873.90 and the partial correlation coefficients are 0.0276 and 0.1308, respectively.It indicates that the

Controlling firm fixed effect
We further control the firm fixed effect to make the conclusion more robust.Table 7 reports the empirical results.The coefficients between DLZZ i,t-1 (GLZZ i,t-1 ) and GAP i,t-1 (GAPD i,t-1 ) have not changed substantially compared with the above empirical analysis.It indicates our results are robust.

Heterogeneity analysis
To obtain certain government subsidies, financing facilities and reputation returns, enterprises can actively assume social responsibility and devote themselves to economic recovery in the economic contraction cycle.It may be affected by the nature of property  rights and individual factors of enterprises.Therefore, we further analyse the impact of the nature of property rights, public popularity, growth level and financial distress on the above relationship.The state-owned enterprises have natural property rights advantages.They are more likely to establish close relationships with the government and have easier access to resources.The non-state-owned enterprises need to cater to the government through rent-seeking or other methods to establish relationships with the government (Kornai et al., 2003).The existing literature indicates that private enterprises can obtain the better market reputations and establish relationships with the government through asset donations, reducing tax avoidance during the government transition period, and then obtaining more resources (Su & He, 2010;Yu et al., 2015).Therefore, corporates with different ownership have different social responsibility motivations in the economic contraction cycle.Ownership may affect the countercyclical support effect of corporate social responsibility.Specifically, compared with state-owned enterprises, non-state-owned enterprises have a weak relationship with government, higher financing constraints and marketisation (Wang et al., 2009).Therefore, non-state-owned enterprises are more motivated to obtain government subsidies, reputation returns and financing convenience by actively fulfilling their social responsibilities.Their countercyclical support effect for society is also stronger.
Enterprises can enhance social responsibility in the economic contraction cycle to obtain government subsidies, financing facilities and reputation returns based on the value-creation effect of social responsibility.Public popularity can promote the countercyclical effect of corporate social responsibility in an economic contraction cycle from the perspective of reputation.Reputation reflects the excellent ability and good reputation of a company or product.It takes a long time to establish development, which is difficult to establish, but easy to destroy (Hall, 1993).Enterprises with higher public popularity have a better reputation.At this time, they are likely to pay more attention to the evaluation of consumers and the development of their reputation in the economic contraction cycle.Enterprises with higher public popularity have stronger motivation to obtain reputation returns.Therefore, we predict that public popularity can promote the motivation of corporate social responsibility in the economic contraction cycle.
Growth level and financial distress can promote the support effect of corporate social responsibility in the economic contraction cycle from the perspective of corporate financing motivation.Generally, the reasons for enterprises to finance are as follows.First, enterprises face attractive investment opportunities.Secondly, the enterprise cash flow is lacking (Wu et al., 2015).The enterprises with higher growth levels are in the stage of rapid development and need a lot of funds to maintain their own development needs.The enterprises with higher growth levels have greater demand for funds and stronger financing motivation.Moreover, enterprises with higher financial difficulty also need to obtain more resources to alleviate their own difficulty.Therefore, we predict that enterprises with higher growth levels and higher financial difficulty have stronger motivation to obtain government subsidies, financing facilities and reputation returns.These enterprises' countercyclical support effect on society is also stronger.
Based on the above analysis, we believe that the countercyclical support effect of corporate social responsibility is significantly stronger in non-state-owned enterprises.
Public popularity, growth level and financial difficulty can significantly promote the above effects.Therefore, we construct Model (2) on the basis of Model (1).
ADJUST i,t-1 is a situational adjustment variable lagging one period, including non-stateowned enterprises(SOE i,t-1 ), public popularity(PUPO i,t-1 ), growth level(GROWD i,t-1 ) and financial distress(ZSCO i,t-1 ).If the enterprise is a non-state-owned enterprise, the SOE i,t-1 is defined as 1, otherwise 0. PUPO i,t-1 is equal to the natural logarithm of the number of financial news.If the growth rate of the main business income of enterprises is greater than the median, GROWD i,t-1 is defined as 1, indicating that the growth ability is high, otherwise it is 0. GROWD i,t-1 is equal to the natural logarithm of the number of financial news.If the Zscore index is less than the median, ZSCO i,t-1 is defined as 1, indicating that the business is in high financial distress, otherwise it is 0. We calculate the Zscore index referring to Altman's method.The larger the Zscore index, the better the financial situation of the enterprise.But the grey interval[1.8,2.675] is defined for American manufacturing companies and may not apply to the Chinese market.So we divide it by median.We decentralise the interactive continuous variables in the above model.
Tables 8-11 report the results of heterogeneity analysis.Table 8 reports the impact of ownership on the relationship between economic contraction cycle and corporate social responsibility.The coefficient of GAPD i,t-1 ×SOE i,t-1 and CSR1 i,t is negative, which may be due to the change in data structure.The regression results are not robust.But the coefficients of the other three columns of GAP i,t-1 ×SOE i,t-1 and GAPD i,t-1 ×SOE i,t-1 are significantly positive.The above results indicate that non-state-owned enterprises have a higher level of social responsibility in the economic contraction cycle.Table 9 reports the impact of public popularity on the relationship between economic contraction cycle and corporate social responsibility.The coefficients of GAP i,t-1 ×PUPO i,t-1 and GAPD i,t-1 ×PUPO i,t-1 are significantly positive.The results indicate that the higher the public popularity, the higher corporate social responsibility in the economic contraction cycle.Table 10 reports the impact of growth level on the relationship between economic contraction cycle and corporate social responsibility.The coefficients of GAP i,t-1 ×PUPO i,t-1 and GAPD i,t-1 ×PUPO i,t-1 are significantly positive.The results indicate that the higher corporate growth level, the stronger the countercyclical support effect of corporate social responsibility.Table 11 reports the impact of financial distress on the relationship between economic contraction cycle and corporate social responsibility.The coefficients of GAP i,t-1 ×ZSCO i,t-1 and GAPD i,t-1 ×ZSCO i,t-1 are significantly positive, indicating that the higher corporate financial distress, the stronger the countercyclical support effect of corporate social responsibility.

Economic consequences
The reason why enterprises play the role of social citizens and actively assume corporate social responsibility for economic recovery may be to obtain subsequent government subsidies, financing facilities and reputational returns in the economic contraction cycle.So do companies get these resources?Do the government and the market compensate for the enterprises' countercyclical support of social responsibility?To verify this problem, we intend to analyse the impact of corporate social responsibility in the economic contraction cycle on government subsidies, financing convenience and reputation returns in the next period.
We construct the model(3) on the basic model(1).
SUBS i,t+1 represents the government subsidy of the t + 1 period.It is equal to the natural logarithm of the enterprise government subsidy and its mean is 15.6874.REPU i,t+1 represents the reputation return of the t + 1 period.It is equal to the natural logarithm of the number of positive news reported by the enterprise and its mean is 4.4006.SA(WW) i,t+1 represents the degree of financing constraints in the t + 1 period.We use the SA index and the WW index to measure it.The SA index is equal to the absolute value of (−0.737*Size +0.043*Size 2 -0.04*Listy) and its mean is 3.7949.The higher SA index, the greater the corporate financing constraint (Ju et al., 2013).The WW index is equal to the absolute value of (−0.091*CF-0.062*DivPos+0.021*Lev-0.044*Size+0.102*ISG-0.035*SG) and its mean is −1.0237.The higher WW index, the greater corporate financing constraint (Liu et al., 2015).
Table 12 and 13 reports the empirical results.The coefficients of CSR i,t in column (1) of Table 12 are significantly positive.It indicates that enterprises in the economic contraction cycle to assume social responsibility will receive government subsidies in a future period.The coefficients of CSR i,t in column (2) of Table 12 are significantly positive.It indicates that enterprises in the economic contraction cycle to assume social responsibility will receive the reputation return in a future period.The coefficients of CSR i,t in Table 13 are significantly negative.It indicates that enterprises in the economic contraction cycle to assume social responsibility will receive financing convenience in a future period.

Conclusions and suggestions
In the critical period of economic transformation, understanding the motivation of corporate behaviour and supervising corporate social responsibility is crucial for promoting social welfare and high-quality economic development.We take A-share listed companies in China's capital market from 2011 to 2018 as research samples, and empirically analyse the relationship between the economic cycle and corporate social responsibility.We further study the role of situational factors such as ownership, public popularity, growth level and financial difficulty, and analyse the economic consequences of corporate social responsibility under the economic  contraction cycle from the perspectives of government, financial institutions and consumers.Our study shows that enterprises have better social responsibility performance in the economic contraction cycle.Corporate social responsibility has a significant countercyclical support effect and the support effect is more obvious in non-state-owned enterprises.Public popularity, growth level and financial difficulties can significantly promote the countercyclical support effect.
Our study not only expands the research on economic cycle and corporate social responsibility, but also helps society to understand macroeconomic development and helps government to supervise corporate social responsibility.It eventually can promote high-quality and sustainable development of society.We propose the following three policy recommendations.First, we should pay attention to the change in corporate social responsibility performance motivation in the economic contraction cycle to ensure the long-term sustainable development of enterprises and society.In China's special political background, enterprises can enhance social responsibility performance to obtain government subsidies, financing facilities and reputation returns in the economic contraction cycle based on the value creation effect.Corporate social responsibility has a significant counter-cyclical support effect.Therefore, we should take into account the motivation of enterprises in the economic contraction cycle in supervision and performance of social responsibility.It can ensure the rational allocation of resources and sustainable economic development.Second, the government should formulate countercyclical policies and regulations about social responsibility.Enterprises can actively enhance social responsibility performance in the contraction cycle based on government resources, financing convenience and reputation return motivation, showing certain support for economic recovery.In the current critical period of macroeconomic slowdown and structural transformation, the government and relevant supervisory departments should strengthen the policy guidance on the implementation of corporate social responsibility under the economic contraction cycle.Meanwhile, when formulating or implementing relevant countercyclical regulations and measures, they should fully consider the changes in the motives for fulfilling corporate social responsibilities under the economic contraction cycle, and accurately understand the regulatory power and flexibility for counter-cyclicality to ensure optimal allocation of resources and promote balanced economic growth.Third, we should pay attention to the effect of situational factors such as ownship, public popularity, growth level and financial difficulty on the countercyclical support effect of social responsibility.The countercyclical support effect of corporate social responsibility is more obvious in non-state-owned enterprises.Public awareness, growth level and financial difficulty can significantly promote the countercyclical support effect of corporate social responsibility.Therefore, enterprises and governments should fully consider the influence of the above situational factors when conducting self-resource regulation and government guidance.Then they could keep the flexibility of optimising resource allocation to ensure stable economic development.

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

Funding
Zejiang Zhou appreciate financial support from the National Natural Science Foundation of China (approval number 72172001) and Anhui Provincial Natural Science Fund Project(approval number 2208085Y22).

Table 3 .
Economic tightening cycle and corporate social responsibility.
***, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level.

Table 4 .
Robustness test: Using new measure method about variable.**, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level. *

Table 5 .
Robustness test: The first stage of instrumental variable method.
***, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level.model is no problem with insufficient identification of instrumental variables.The values of Score chi 2 are 3.3511, 2.0672, 0.8525 and 1.8087, respectively, which are not statistically significant.It indicates that the model is no problem with over-identifying of instrumental variables.The above results indicate that the research conclusions are not basically affected by endogenous problems.

Table 6 .
Robustness test: The second stage of instrumental variable method.**, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level. *

Table 7 .
Robustness test: Controlling fixed effect of firm.
***, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level.

Table 8 .
Heterogeneity analysis: The moderating effect of SOE.**, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level. *

Table 9 .
Heterogeneity analysis: The moderating effect of public popularity.**, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level. *

Table 10 .
Heterogeneity analysis: The moderating effect of growth level.**, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level. *

Table 11 .
Heterogeneity analysis: The moderating effect of financial distress.**, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level. *

Table 12 .
Social responsibility and government subsidy(reputation) during business contraction.Note: ***, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level.

Table 13 .
Social responsibility and financing constraint during business contraction.**, **, and * represent the 1%, 5%, and 10% levels of significance, respectively, for a two-tailed test.All reported T-values and Z-values in parentheses are based on standard errors adjusted for clustering at the firm level. *