How Corporate Social Responsibility Affects Firm Performance: The Inverted-U Shape Contingent on Founder CEO

Abstract

Despite abundant research on the relationship between CSR and firm performance, prior research generated highly inconsistent findings. No consensus has been achieved on the relationship between CSR and firm performance. The objective of this research is to examine how the relationship between CSR and firm performance is contingent on founders’ management roles, especially in the situation of the founder as CEO, which will provide insights into the inconsistent impacts of CSR. Based on panel data analysis, we empirically test the nonlinear relationship between corporate social responsibility (CSR) and firm performance for China’s Growth Enterprise Market (GEM) listed companies. We further explore how this relationship differs under two types of CEOs: founder CEOs and non-founder CEOs. Our econometric analysis produces two major findings. First, there is an inverted U-shaped relationship between CSR and firm performance. Second, the presence of founder CEO weakens the relationship between CSR and firm performance, making the inverted U-shaped curve flatter. This research makes both novel theoretical and practical contributions to entrepreneurship and organization research by providing an enriched understanding of the relationship between CSR and firm performance. It integrates multiple theories to create a framework within which the contingent impacts of CSR can be holistically understood. It also helps managers to realize the nonlinear economic consequences of CSR activities and the different regulatory effects of founder management.

1. Introduction

Corporate social responsibility (CSR) reflects a corporation’s responsibility to society and a wider range of stakeholders other than its shareholders [1], which has drawn tremendous attention from researchers and managers due to its impact on corporate development. CSR has been found to be able to improve corporate performance [2,3,4,5], enhance corporate reputation [6,7,8], promote corporate innovation [9,10], reduce corporate risks [11,12], and maintain corporate stability in the face of emergencies to achieve economic recovery [13]. However, Wright and Ferris [14] found that CSR could damage enterprise value. Under different levels of CSR, the balance between benefits and costs is constantly changing and has different impacts on the enterprise. A simple linear relationship cannot precisely describe the specific economic consequences of CSR [15], as excessive CSR investments may produce negative effects on a corporation. Existing research has demonstrated that CSR has inverted U-shaped economic consequences on various outcomes, such as shareholder value [16] and corporate risk [17]. Additionally, human resource management, as an important factor affecting CSR, has also been shown to have an inverted U-shaped relationship with both CSR and financial performance [18]. In the context of start-up enterprises, we focus on the impact of CSR on enterprise performance to help enterprises make better CSR investments. We believe that CSR will have an inverted U-shaped impact on corporate performance, which provides a rich understanding of the inverted U-shaped economic consequences of CSR.

Despite abundant research on the relationship between CSR and firm performance, prior research generated highly inconsistent findings. This relationship is found to be positive [5,19,20,21], negative [22,23,24,25], nonlinear [26,27,28], or non-significant [29,30] (see Appendix A for a review of previous studies). No consensus has been achieved on the relationship between CSR and firm performance. Isaksson and Woodside [31] suggest that the relationship between CSR and firm performance would be affected by corporate governance, and corporate governance factors should be taken into account in the analysis. Servaes and Tamayo [32] and Crifo, Diaye, and Pekovic [33] point out that corporate governance can potentially moderate the relationship between CSR and firm performance. As the core figure of start-up companies, the founders establish and lead the companies to grow continuously. Their position in the company is an important corporate governance arrangement, which could affect the consequences of CSR. Therefore, it is essential to examine how the relationship between CSR and firm performance is contingent on founders’ management roles, especially in the situation of the founder as CEO, which will provide insights into the inconsistent impacts of CSR.

This paper intends to answer two research questions: (1) What is the relationship between CSR and firm performance? (2) How is the relationship between CSR and firm performance different when the firm is led by a founder CEO or a professional CEO? To answer these questions, we analyze the sample of China’s GEM listed companies from 2010 to 2019 to explore the moderating effect of the founder’s management role on the relationship between CSR and firm performance.

This study makes two major contributions. First, China is the largest emerging economy in the world. Focusing on the specific implementation of CSR of start-up listed companies for China’s GEM and drawing from stakeholder theory and resource dependence theory, we propose and validate the inverted-U shaped relationship between CSR and firm performance to enhance the understanding of the CSR impacts on start-up firms. Second, while prior research shows that management ownership has nonlinear relationships with various outcomes, such as TBQ [34] and carbon transparency [35], management ownership is studied as the independent variable. Different from the prior research, we focus on the moderating role of a specific type of management ownership, founder CEO. Drawing from agency theory and stewardship theory, we find different effects of the founder CEO and the professional CEO on the relationship between CSR and firm performance, integrating both CSR and corporate governance to shed light on entrepreneurship research in small- and medium-sized firms.

2. Theoretical Foundation

Many theories have been employed in the research on CSR in recent decades, which provide a rich foundation for studying the influencing factors and economic consequences of CSR [36]. Following Frynas and Yamahaki [37], we integrate four theories to guide our research on the relationship between CSR and firm performance and the moderating effect of founder CEOs’ role on this relationship. Specifically, stakeholder theory and resource dependence theory are employed to explain the curvilinear relationship between CSR and firm performance, and agency theory and stewardship theory are employed to explain the moderating effect of founder CEO.

2.1. Stakeholder Theory

Stakeholders refer to individuals or groups that may affect or be affected by the corporate, which mainly include employees, customers, suppliers, shareholders, governments, communities, managers, and so on [38,39]. In the view of stakeholder theory, the relationship between the corporate and its stakeholders, especially its main stakeholders, determines corporate survival and growth [40,41]. Stakeholder theory suggests that stakeholder relationships should be considered as an analysis unit to examine how corporates use these relationships to co-create and conduct value transactions [39]. At the same time, it also suggests that focusing on the creation and maintenance of stakeholder relationships can effectively improve corporate value creation [42].

Barnett [43] suggests that CSR can create value because it helps to generate positive stakeholder relationships. However, Barnea and Rubin [44] suggest that CSR will cause conflicts among shareholders and reduce CSR’s contribution to corporate value, because there is no limit to the amount that enterprises transfer to stakeholders for CSR investment, the marginal effect of extra CSR expenditure per unit will inevitably reduce shareholders’ wealth. This research found that excessive investment in CSR may reduce corporate value.

2.2. Resource Dependence Theory

Resource refers to everything that can be considered to give corporate advantages or disadvantages [45]. According to resource dependence theory, the key resources needed by a corporation are controlled by stakeholders, and the corporation can gain support from stakeholders by obtaining key resources to increase the efficiency of resource use [29] and ultimately improve corporate performance [19]. For example, studies have shown that enterprises that pay attention to CSR will give employees a greater sense of belonging [46]. Enterprises that pay attention to CSR in terms of products or services may obtain higher product demand [47] because customers are more likely to pay a premium for such enterprises [48]. Moreover, CSR can improve the relationship with governments to reduce the impact of government regulatory interventions [49]. These will help enterprises improve their competitive position in the industry, and directly or indirectly contribute to shareholders’ value creation and sustained corporate competitive advantages [50].

However, to meet the social goals in CSR and form a good stakeholder relationship, enterprises will invest lots of resources in CSR practice, forming “extra costs” for the corporation [47]. There was evidence that not all enterprises can afford the “extra costs” of CSR [20]. For example, Wang and Bansal [51] suggest that resource constraints are common in new enterprises, which will hinder enterprises from incorporating social and environmental innovation into product features and organizational processes. Their empirical results show that due to the limited resources of new enterprises, CSR activities and firm performance are negatively correlated. Under resource scarcity, more resources invested in CSR mean fewer resources available to develop the core corporate competitiveness [47,52], which may lower firm performance.

2.3. Agency Theory

Agency theory analyzes how the “principal” (the individual or organization that hires the other party to perform a specific job) and the “agent” (the person hired to do the job) are related [37]. It explains interest incongruence between the principal and the agent, the agency problem caused by the agent’s private behavior, and helps to solve this agency problem through appropriate corporate governance methods [53,54,55].

With regard to the application of agency theory in the field of CSR, Friedman [56] pioneered and suggested that CSR was a selfish behavior of corporate managers as agents. Their realization of social goals will ultimately lead to lower corporate profits and damage the interests of shareholders as principals. Other scholars endorse this view: for example, Wang, Feng and Palmer [20] suggest that CSR is a kind of executive benefit as managers use CSR to pursue personal interests or promote their own professional development. Barnea and Rubin [44] find that corporate managers have a tendency to over-invest in CSR in order to seek private interests and improve their personal reputation. From this angle, CSR is seen as an abuse of corporate resources, assuming that corporate resources should be better used for internal projects to generate value [57].

2.4. Stewardship Theory

Based on psychology and sociology research, stewardship theory examines the situation where the managerial motivation of managers (agents) is to maximize the interests of their principals [58]. Different from agency theory, stewardship theory suggests that the goals between managers and their principals tended to be consistent, guiding the company and its shareholders to conduct trustworthy ethical behavior [59].

Under stewardship theory, managers play the role of stewards and can produce ethical corporate behaviors. Godos-Diez, Fernandez-Gago and Martinez-Campillo [60] surveyed 149 CEOs in Spain and found that CEOs showing more stewardship would pay more attention to corporate ethics and CSR, and could promote higher level CSR practice. Lamb and Butler [61] found that managers with housekeeping roles in family enterprises are more likely to attach importance to CSR in pursuit of long-term organizational goals through comparative analysis of samples of family enterprises and non-family enterprises.

3. Hypotheses Development

3.1. CSR and Firm Performance

It has been widely recognized that CSR can promote organizational performance, but the relationship between CSR and firm performance may be too complex to be fully explained by a simple linear relationship [62,63]. We propose a nonlinear relationship between CSR and firm performance based on the following reasonings.

Based on the stakeholder theory, some enterprises implement CSR because of accountability to customers, the public, and other parties, so as to meet the stakeholders’ needs [25]. In the process of meeting the needs of stakeholders, CSR improves the corporate reputation in society and citizens, attracts the attention and support of stakeholders [64], provides similar “insurance-style” protection for corporate assets [65,66], and enables the corporation to obtain key resources from stakeholders [67]. Related professional skills or abilities gained from CSR activities can bring specific economic benefits to the enterprise [50]. For example, CSR can increase the enterprise’s intangible assets, strengthen the relationship with stakeholders, bring competitive advantages, and improve the firm performance [68]. These research results show that CSR can have a positive impact on firm performance.

However, different CSR activities influence business outcomes and subsequent firm performance through different stakeholder value propositions [69]. Response of stakeholders to CSR behavior changes at different CSR levels. Therefore, enterprises with different CSR levels will receive different CSR returns [27]. Brammer and Millington [70] point out that the stakeholders’ awareness of CSR is a utility function with diminishing marginal utility. They show that CSR’s effect on improving the relationship between stakeholders can bring positive economic benefits, but ultimately these benefits will be diminishing. Moreover, CSR plans or activities are often separated from the corporate core business or have nothing to do with its shareholder value, which may reduce its contribution to firm performance [71]. Those studies show that the impact of CSR activities closely related to stakeholders on firm performance will change with the level of CSR, and over-investment in CSR may reduce firm performance.

If the CSR investment is relatively low, enterprises can profit from it, but the marginal benefits will gradually decline as the CSR investment grows. Continuously increasing the amount of CSR investment may eventually lead to a decline in revenue. The reason for this interesting phenomenon can be understood from the perspective of resource dependence theory. First, the resources obtained from stakeholders are limited, both in terms of quantity and type, these restrictions inhibit the corporation’s continuing profitability [66]. Second, even if stakeholders can continuously provide a steady stream of resources for the enterprise, it is impossible to maintain linear growth in revenue. Stakeholders expect to get a reasonable return from the enterprise when they give the enterprise resources. With the continuous increase of CSR, stakeholders will reduce their resource provision because the marginal benefit of stakeholders’ resource giving will decrease as the level of giving increases [66]. When an enterprise continuously increases CSR, it will inevitably increase product prices, reduce employee salaries, or reduce investor returns, and so on, to transfer part of the CSR costs to stakeholders [47]. Font and Lynes [28] argue that CSR has a positive impact on firm performance, but the effect may follow an inverted U-shaped trend. Thus, the impact of CSR activities on firm performance will change with the level of CSR, and over-investment in CSR may reduce firm performance.

In summary, a corporation can establish a good corporate image, improve corporate reputation, increase corporate competitiveness, bring economic benefits to the corporation, and improve firm performance by fulfilling CSR that is under a certain threshold. The different responses of stakeholders to different levels of CSR affect the level of profit that an enterprise can make from stakeholders through CSR activities. With the continuous investment in CSR, the benefits of CSR activities by enterprises will gradually saturate. Excessive investment in CSR leads to a downward trend in firm performance, as the benefits obtained from stakeholders are offset by increasing CSR expenses. The relationship between CSR and firm performance can be summarized as follows: CSR can have a positive impact on firm performance before a threshold level, and excessive CSR investment over the threshold can produce negative firm returns. In the process of increasing CSR investment, the impact on firm performance increases first and then decreases, producing an inverted U-shaped curve. Thus, the following hypotheses are proposed:

Hypothesis H1 (H1).

There is an inverted U-shaped relationship between CSR and firm performance.

3.2. Moderating Effects of the Founder Management Role

According to agency theory, the CEO as an agent has greater management power [72] in determining CSR activities, so it can be regarded as a private CSR agent [73]. Cho and Lee [74] argue that the CEO is facing the dilemma of CSR investment when dealing with the fierce conflict and competition between different stakeholders. However, for founder CEOs, their role switched from an agent to a steward. In another word, the agent role and steward role converge when the founder serves as the CEO. Compared with non-founder CEOs, founder CEOs are more devoted to the success of their enterprises and consider their enterprise as a proud accomplishment [75]. The pride brought by enterprise success or the stigma brought by enterprise failure is much stronger for them compared with non-founder CEOs [46,76]. Thus, we contend that non-founder CEOs’ role is characterized by agency behaviors, whereas founder CEOs’ role is characterized by stewardship behaviors and this intrinsic difference between CEOs could moderate how firms’ CSR influence their firm performance.

First, at the low to medium levels of CSR, while CSR can quickly attract resources for the firm, the founder CEO is more conservative and cautious in relying on CSR to improve firm performance because of their stewardship role [77,78]. Compared with non-founder CEOs, founder CEOs may be focused more on the social benefits of CSR and take less aggressive approaches to reap monetary benefits from CSR activities. As a result, for the same number of increases in CSR, firms led by founder CEOs may have less gain in firm performance than firms led by non-founder CEOs.

On the other hand, as the level of CSR increases from medium to high, the firm’s performance will suffer but the CEOs of these firms will continue to gain a positive social image. In the process of continuous CSR investment, founder CEOs concentrate on playing the role of a qualified steward in improving the collective benefits of the firm [79], and are less likely to engage in agency behaviors than non-founder CEOs [75,80]. Compared with non-founder CEOs, founder CEOs are more likely to beware the potential pitfalls of excessive CSR and actively take efforts to make up for CSR costs and alleviate the negative impact of CSR on firm performance. Because founder CEOs have special feelings and stronger emotional ties to their enterprises, it is unlikely that they will use CSR to improve his personal interests at the cost of the enterprise. Therefore, CSR’s negative impact on firm performance is reduced. In contrast, non-founder CEOs are more indifferent to the negative impact of excessive CSR because of their agency attribute. As a result, firm performance will have a steeper decline as CSR increases. Thus, we propose the following hypothesis:

Hypothesis H2 (H2).

The presence of founder CEO moderates the relationship between CSR and firm performance, so that the inverted-U-shaped effect of CSR on firm performance is more flattened for founder CEOs than non-founder CEOs.

Our research model is shown in Figure 1.

4. Method

4.1. China’s GEM

Growth Enterprise Market (GEM) is the second stock exchange market in China. It is a trading market designed to provide financing channels and growth space for entrepreneurial enterprises that are unable to be listed on the Main Board Market. As an important supplementary part of the Main Board Market, it is the cradle for cultivating private enterprises and plays an important role in the capital market in China [81].

With the high-speed economic development in recent years, Chinese firms have gradually realized the importance of fulfilling social responsibilities. According to a 2019 survey by the Institute of Finance and Banking, Chinese Academy of Social Sciences, among listed companies on the Main Board Market, Small and Medium-sized Board Market and GEM, those listed on GEM had the highest average score of comprehensive quality (66.27) but the lowest average score of social responsibility (41.48). (Data from Institute of Finance and Banking, Chinese Academy of Social Sciences: Quality Evaluation Report of Chinese Listed Companies from 2018 to 2019).

4.2. Sample

Our sample is China’s GEM listed companies from 2010 to 2019. The CSR data are obtained from the social responsibility report of listed companies published by Hexun.com (accessed on 16 May 2022). (Hexun.com (accessed on 16 May 2022) releases the social responsibility report of listed companies every year to evaluate the performance of social responsibility of Listed companies in China). The other data are extracted from the China Stock Market & Accounting Research (CSMAR) database. In order to unify the sample data and avoid the disturbance of outliers in the company’s financial condition, we preprocessed the data as follows: (1) Companies with ST and *ST (ST refers to the situation in which a company has suffered losses for two consecutive years and has been specially treated. *ST refers to in the third year, the company’s operations have not improved and it is still in a state of loss. In addition to the “ST” before the stock name, “*”will be added, which means the risk of delisting.) in the time interval were removed; (2) Observations with missing values in the data were removed; (3) All continuous variable data were winsorized (1% and 99% levels). Our final data contains 731 companies with 4745 observations for a decade. This data is an unbalanced panel data because the time for the companies to be listed on the GEM is not consistent.

4.3. Variables

Dependent variable. Following prior research, we measure firm performance by using market values (MV) [82].

Independent variable. Corporate social responsibility (CSR) is measured by a composite score reported by Hexun.com (accessed on 16 May 2022), a well-known finance and economics platform in China. We divide the CSR scores by 100 for standardization. (The maximum of CSR in Hexun’s listed company social responsibility report is 100.) The standardized CSR scores were used to generate a squared term (CSR²) to test the quadratic relationship between CSR and firm performance [83].

Moderators. The moderating variable is the presence of founder CEO ($Fce{o}_{i,t}$), if the CEO of company i is a founder in year t, $Fce{o}_{i,t}$ is coded as 1, otherwise coded as 0 [84].

Control variables. Based on previous studies [85], we control for the effects of board size, independent directors, asset-liability ratio, enterprise size, enterprise age, enterprise growth and degree of ownership concentration. In addition, the fixed effects of year and industry are controlled.

Table 1. Description of variables.

Variables	Measurement	Source
Market values (MV)	Logarithm of the market values of the company’s shares	CSMAR
Corporate social responsibility (CSR)	The social responsibility score report published by Hexun.com (accessed on 16 May 2022), divided by 100 for standardization	Hexun.com (accessed on 16 May 2022)
Founder CEO (FCEO)	If the founder or one of the founder team members serves as CEO, coded as 1, otherwise coded as 0	CSMAR
Board size	The actual number of corporate board of directors	CSMAR
Independent directors	The number of independent directors in the corporate board of directors	CSMAR
Asset-liability ratio	Total liabilities/Total assets	CSMAR
Enterprise size	Logarithm of number of total employees	CSMAR
Enterprise age	Logarithm of number of years since the firm first listed	CSMAR
Enterprise growth	The growth rate of operating revenue	CSMAR
Ownership concentration	The ratio of the total number of shares held by the largest shareholder divided by the total share capital of the company	CSMAR
Industry	Industry FE	CSMAR
Year	Year FE	CSMAR

shows the variables and their definitions.

4.4. Empirical Models

We use Stata15.1 for data analysis. To test H1, we use a two-way fixed effects model to determine the relationship between CSR and firm performance. In Model 1, $CS{R}_{i,t}$ and $CS{R}_{i,t}^2$ denote the main effect and quadratic effect of CSR. Year and industry fixed effects are controlled for to address unexplained variance clustered within time and industry.

$$M{V}_{i,t}{=\alpha }_0{+\alpha }_{1}CS{R}_{i, t}{+\alpha }_{2}CS{R}_{i,t}^2+{{\displaystyle \sum }}^{ }Control{s}_{i,t}+{{\displaystyle \sum }}^{ }{\mathit{Industry}}_i+{{\displaystyle \sum }}^{ }{\mathit{Year}}_t{+\epsilon }_{i,t}$$

(1)

where $i$ indexes firm and t indexes year $t\in \left(2010, 2019\right)$.

To test H2, Models 2 and 3 are estimated in which $M_{i,t}$ refers to ${\mathit{Fceo}}_{i,t}$. Note that ${\mathit{Fceo}}_{i,t}$ is a time-varying variable. In Model 3, ${\theta }_5$ represents the slope change of the inverted U-shaped curve, which is the key to understanding the moderating effect under the difference of the founder CEO role [63].

$$M{V}_{i,t}={\eta }_0{+\eta }_{1}CS{R}_{i, t}{+\eta }_{2}CS{R}_{i,t}^2{+\eta }_3{M}_{i,t}+{{\displaystyle \sum }}^{ }{\mathit{Controls}}_{i,t}+{{\displaystyle \sum }}^{ }{\mathit{Industry}}_i+{{\displaystyle \sum }}^{ }{\mathit{Year}}_t{+\epsilon }_{i,t}$$

(2)

$$\begin{array}{c}M{V}_{i,t}={\theta }_0+{\theta }_{1}CS{R}_{i, t}+{ \theta }_{2}CS{R}_{i,t}^2+{\theta }_3{M}_{i,t}+{\theta }_4{M}_{i,t}\ast CS{R}_{i,t}+{\theta }_5{M}_{i,t}\ast CS{R}_{i,t}^2\\ +{{\displaystyle \sum }}^{ }{\mathit{Controls}}_{i,t}+{{\displaystyle \sum }}^{ }{\mathit{Industry}}_i+{{\displaystyle \sum }}^{ }{\mathit{Year}}_t{+\epsilon }_{i,t}\end{array}$$

(3)

5. Results

5.1. Descriptive Statistics

Table 2. Descriptive statistics.

Variables	N	Mean	SD	Min	Max
Market values (MV)	4745	20.75	5.093	0	24.333
Corporate social responsibility (CSR)	4745	0.214	0.106	−0.036	0.672
Founder CEO (FCEO)	4745	0.540	0.498	0.000	1.000
Board size	4745	7.982	1.432	5.000	12.000
Independent directors	4745	2.993	0.389	2.000	4.000
Asset-liability ratio	4745	0.298	0.175	0.033	0.787
Enterprise size	4745	6.861	0.859	4.990	9.202
Enterprise age	4745	2.619	0.381	1.245	3.298
Enterprise growth	4745	0.208	0.367	−0.489	1.877
Degree of ownership concentration	4745	0.311	0.125	0.089	0.639

presents the descriptive statistics of all the variables. The mean of CSR ($CS{R}_{i,t}$) is 0.214, indicating that the overall CSR performance level of the sample companies is generally low, the minimum value is −0.036, and the maximum value is 0.672, which shows that there are great differences in CSR performance. The market values ($M{V}_{i,t}$) is 20.75, the minimum value is close to 0, and the maximum value is 24.333, indicating that the overall market value of the sample companies is relatively high, but at the same time, there is also a clear degree of distinction between the samples. The mean of founder CEO is 0.540, indicating that 54% CEOs in the sample enterprises are held by founders.

Table 3. Correlations between variables.

Variables	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.
1. Market values (MV)	1
2. CSR	0.131 ***	1
3. Founder CEO	0.050 ***	0.072 ***	1
4. Board size	0.031 **	0.060 ***	−0.014	1
5. Independent directors	0.010	0.033 **	0.030 **	0.666 ***	1
6. Asset-liability ratio	−0.070 ***	−0.280 ***	−0.089 ***	0.002	0.009	1
7. Enterprise size	0.022	−0.008	−0.071 ***	0.105 ***	0.106 ***	0.342 ***	1
8. Enterprise age	−0.120 ***	−0.112 ***	−0.028 *	−0.003	0.002	0.152 ***	0.117 ***	1
9. Enterprise growth	0.110 ***	0.167 ***	−0.016	0.031 **	0.019	0.118 ***	0.177 ***	−0.017	1
10. Degree of ownership concentration	0.057 ***	0.084 ***	0.043 ***	−0.118 ***	−0.071 ***	−0.063 ***	−0.090 ***	−0.070 ***	−0.023	1

*** p < 0.01, ** p < 0.05, * p < 0.1, one-tailed test.

shows correlations among all variables. It is found that the coefficients between variables are all less than 0.7, showing no multicollinearity problem.

5.2. Effects of CSR on Firm Performance

The Hausman test shows that the fixed effects (FE) model is more appropriate than the random effects model. Therefore, we employ the FE estimator in all the regression analyses. As shown in

Table 4. Regression results of CSR and firm performance.

Variables	Model 1	Model 2	Model 3
CSR	5.292 ***	5.200 ***	8.026 ***
	(1.809)	(1.809)	(2.324)
CSR2	−7.271 ***	−7.102 **	−12.462 ***
	(2.808)	(2.808)	(3.752)
FCEO		0.450 *	1.099 **
		(0.233)	(0.479)
CSR × FCEO			−6.183 *
			(3.195)
CSR2 × FCEO			11.282 **
			(5.227)
Board size	0.005	0.016	0.004
	(0.110)	(0.110)	(0.111)
Independent directors	−0.057	−0.073	−0.066
	(0.357)	(0.357)	(0.357)
Asset-liability ratio	1.129	1.148	1.196
	(0.769)	(0.769)	(0.770)
Enterprise size	0.176	0.170	0.167
	(0.194)	(0.194)	(0.194)
Enterprise age	−0.014	−0.026	0.013
	(0.971)	(0.971)	(0.972)
Enterprise growth	0.841 ***	0.846 ***	0.846 ***
	(0.199)	(0.199)	(0.199)
Degree of ownership concentration	2.852 *	2.768*	2.697 *
	(1.605)	(1.605)	(1.607)
Cash holdings	Yes	Yes	Yes
	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes
Year FE	36.344 ***	36.214 ***	36.165 ***
Firm FE	(3.770)	(3.769)	(3.769)
Constant	4745	4745	4745
	0.285	0.286	0.287
Observations	5.292 ***	5.200 ***	8.026 ***
R-squared	(1.809)	(1.809)	(2.324)

Note: Standard errors are in brackets; the significance level are *** p < 0.01, ** p < 0.05, * p < 0.1, one-tailed test.

, the main effect of CSR is positive (α = 5.292, p < 0.01), but the quadratic effect of CSR ($CS{R}_{i,t}^2$) is negative (α = −7.271, p < 0.01), supporting H1. This suggests an inverted U-shaped relationship between CSR and firm performance and choosing the appropriate CSR investment level conduce to maximizing firm performance.

As Figure 2 shows, CSR investment first improves firm performance until reaching the optimal CSR input level at which firm performance is the highest. After that, firm performance starts to decline as CSR increases. At the turning point, CSR is 0.364, indicating that the turning point exists within the range of valid CSR values of our sample (min = −0.036, max = 0.672), thus confirming the existence of an inverted-U shaped curve.

5.3. Moderation Effect of Founder CEO

As shown in column 3 of Table 4, founder CEO interacts with both CSR ($CS{R}_{i,t}\ast Fce{o}_{i,t}$) (α = −6.183, p < 0.10) and its quadric term ($CS{R}_{i,t}^2\ast Fce{o}_{i,t}$) (α = 11.282, p < 0.05), supporting H2. This indicates that when the founder is the CEO, the relationship between CSR and firm performance is flattened. As the simple slope illustration shows (Figure 3), the relationship between CSR and firm performance becomes almost linear.

5.4. Robustness Tests

5.4.1. Alternative Dependent and Control Variables

We replaced the dependent variable, MV is replaced with return on assets (ROA) and return on net assets (ROE) [86] and ran the model again. As shown in column 1 of

Table 5. Results with ROA/ROE as the dependent variable.

Variables	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
	ROAt	ROAt	ROAt	ROEt	ROEt	ROEt
CSR	1.039 ***	1.038 ***	1.074 ***	1.935 ***	1.934 ***	2.058 ***
	(0.019)	(0.019)	(0.024)	(0.040)	(0.040)	(0.052)
CSR2	−1.245 ***	−1.243 ***	−1.289 ***	−2.367 ***	−2.367 ***	−2.528 ***
	(0.029)	(0.029)	(0.039)	(0.062)	(0.062)	(0.083)
FCEO		0.005 **	0.016 ***		0.001	0.037 ***
		(0.002)	(0.005)		(0.005)	(0.011)
CSR × FCEO			−0.080 **			−0.271 ***
			(0.033)			(0.071)
CSR2 × FCEO			0.100 *			0.356 ***
			(0.054)			(0.116)
Board size	0.002	0.002	0.002	0.004 *	0.004 *	0.004
	(0.001)	(0.001)	(0.001)	(0.002)	(0.002)	(0.002)
Independent directors	0.005	0.005	0.005	0.008	0.008	0.008
	(0.004)	(0.004)	(0.004)	(0.008)	(0.008)	(0.008)
Asset-liability ratio	−0.056 ***	−0.055 ***	−0.054 ***	−0.085 ***	−0.085 ***	−0.081 ***
	(0.008)	(0.008)	(0.008)	(0.017)	(0.017)	(0.017)
Enterprise size	0.008 ***	0.008 ***	0.008 ***	0.016 ***	0.016 ***	0.015 ***
	(0.002)	(0.002)	(0.002)	(0.004)	(0.004)	(0.004)
Enterprise age	0.004	0.004	0.005	−0.002	−0.002	0.001
	(0.010)	(0.010)	(0.010)	(0.022)	(0.022)	(0.022)
Enterprise growth	0.016 ***	0.016 ***	0.016 ***	0.034 ***	0.034 ***	0.034 ***
	(0.002)	(0.002)	(0.002)	(0.004)	(0.004)	(0.004)
Degree of ownership concentration	0.158 ***	0.157 ***	0.155 ***	0.295 ***	0.295 ***	0.289 ***
	(0.017)	(0.017)	(0.017)	(0.036)	(0.036)	(0.036)
Industry FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes
Constant	−0.165 ***	−0.167 ***	−0.169 ***	−0.344 ***	−0.345 ***	−0.351 ***
	(0.039)	(0.039)	(0.039)	(0.084)	(0.084)	(0.084)
Observations	4745	4745	4745	4745	4745	4745
R-squared	0.606	0.606	0.607	0.523	0.523	0.525

Note: Standard errors are in brackets; the significance level are *** p < 0.01, ** p < 0.05, * p < 0.1, one-tailed test.

shows, the main effect of CSR is positive (α = 1.039, p < 0.01), but the quadratic effect of CSR ($CS{R}_{i,t}^2$) is negative (α = −1.245, p < 0.01), supporting H1. As shown in column 4 of Table 5 shows, the main effect of CSR is positive (α = 1.935, p < 0.01), and the quadratic effect of CSR ($CS{R}_{i,t}^2$) is negative (α = −2.367, p < 0.01), supporting H1. As shown in column 3 of Table 5, founder CEO interacts with both CSR ($CS{R}_{i,t}\ast Fce{o}_{i,t}$) (α = −0.080, p < 0.05) and its quadric term ($CS{R}_{i,t}^2\ast Fce{o}_{i,t}$) (α = 0.100, p < 0.10), supporting H2. As shown in column 6 of Table 5, founder CEO interacts not only with CSR (α = −0.271, p < 0.01) but also with the quadric term of CSR (α = 0.356, p < 0.01), supporting H2.

We measured firm size by using the logarithm of total assets, and included this new firm size as a control variable in the analysis. As

Table 6. Results with ROA/ROE as the dependent variable and controlling for total assets as firm size.

Variables	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
	ROAt	ROAt	ROAt	ROEt	ROEt	ROEt
CSR	1.047 ***	1.046 ***	1.084 ***	1.942 ***	1.942 ***	2.067 ***
	(0.019)	(0.019)	(0.024)	(0.041)	(0.041)	(0.052)
CSR2	−1.254 ***	−1.252 ***	−1.299 ***	−2.375 ***	−2.375 ***	−2.538 ***
	(0.029)	(0.029)	(0.039)	(0.063)	(0.063)	(0.084)
Fceo		0.005 **	0.016 ***		0.002	0.038 ***
		(0.002)	(0.005)		(0.005)	(0.011)
CSR × Fceo			−0.083 **			−0.273 ***
			(0.033)			(0.071)
CSR2 × Fceo			0.103 *			0.357 ***
			(0.054)			(0.116)
Board size	0.002*	0.002*	0.002 *	0.005 **	0.005 **	0.005 *
	(0.001)	(0.001)	(0.001)	(0.002)	(0.002)	(0.002)
Independent directors	0.005	0.005	0.005	0.008	0.008	0.008
	(0.004)	(0.004)	(0.004)	(0.008)	(0.008)	(0.008)
Asset-liability ratio	−0.047 ***	−0.047 ***	−0.046 ***	−0.073 ***	−0.073 ***	−0.069 ***
	(0.008)	(0.008)	(0.008)	(0.017)	(0.017)	(0.017)
Enterprise size	−0.002	−0.002	−0.002	0.003	0.003	0.002
	(0.002)	(0.002)	(0.002)	(0.004)	(0.004)	(0.004)
Enterprise age	0.004	0.003	0.004	−0.002	−0.002	0.001
	(0.010)	(0.010)	(0.010)	(0.022)	(0.022)	(0.022)
Enterprise growth	0.018 ***	0.018 ***	0.018 ***	0.036 ***	0.036 ***	0.036 ***
	(0.002)	(0.002)	(0.00 2)	(0.004)	(0.004)	(0.004)
Degree of ownership concentration	0.147 ***	0.146 ***	0.143 ***	0.289 ***	0.288 ***	0.280 ***
	(0.017)	(0.017)	(0.017)	(0.037)	(0.037)	(0.037)
Industry FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes
Constant	−0.080	−0.081	−0.078	−0.303**	−0.303**	−0.293**
	(0.057)	(0.057)	(0.057)	(0.122)	(0.122)	(0.122)
Observations	4745	4745	4745	4745	4745	4745
R-squared	0.604	0.605	0.605	0.522	0.522	0.523

Note: Standard errors are in brackets; the significance level are *** p < 0.01, ** p < 0.05, * p < 0.1, one-tailed test.

shows, when the dependent variable is ROA, the main effect of CSR is positive (α = 1.047, p < 0.01), but the quadratic effect of CSR is negative (α = −1.254, p < 0.01), supporting H1. The founder CEO interacts not only with CSR (α = −0.083, p < 0.05), but also with its quadric term (α = 0.103, p < 0.10), supporting H2. When the dependent variable is ROE, the main effect of CSR is positive (α = 1.942, p < 0.01), but the quadratic effect of CSR is negative (α = −2.375, p < 0.01), supporting H1. The founder CEO interacts not only with CSR (α = −0.273, p < 0.01) but also with its quadric term (α = 0.357, p < 0.01), supporting H2.

5.4.2. Adding Control Variables

To control for the confounding effects of firms’ profitability and innovation intensity, we added several new control variables into the fixed effects model: sales expense ratio (the ratio of enterprise sales expenses to operating income in year t), income tax expense ratio (the ratio of corporate income tax expenses to operating income in year t), and R&D intensity (the ratio of R&D investment to operating income in year t). The results in

Table 7. Add control variables results.

Variables	Model 1	Model 2	Model 3
	MVt	MVt	MVt
CSR	6.448 ***	6.407 ***	8.965 ***
	(1.925)	(1.925)	(2.398)
CSR2	−8.220 ***	−8.091 ***	−13.145 ***
	(2.878)	(2.878)	(3.789)
FCEO		0.485 **	1.067 **
		(0.233)	(0.479)
CSR × FCEO			−5.707 *
			(3.196)
CSR2 × FCEO			10.704 **
			(5.227)
Board size	0.006	0.018	0.007
	(0.110)	(0.110)	(0.111)
Independent directors	−0.033	−0.049	−0.043
	(0.357)	(0.357)	(0.356)
Asset-liability ratio	1.014	1.037	1.075
	(0.777)	(0.777)	(0.778)
Enterprise size	0.142	0.133	0.132
	(0.195)	(0.195)	(0.195)
Enterprise age	−0.021	−0.033	−0.002
	(0.974)	(0.974)	(0.974)
Enterprise growth	0.799***	0.804 ***	0.804 ***
	(0.202)	(0.202)	(0.202)
Degree of ownership concentration	2.684*	2.585	2.534
	(1.605)	(1.605)	(1.607)
Sales expense ratio	−14.897***	−15.379 ***	−14.997 ***
	(4.951)	(4.955)	(4.960)
Income tax expense ratio	−0.023	−0.024	−0.024
	(0.025)	(0.025)	(0.025)
R&D investment	0.006	0.018	0.007
	(0.110)	(0.110)	(0.111)
Industry FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes
Constant	36.579 ***	36.418 ***	36.398 ***
	(3.791)	(3.790)	(3.791)
Observations	4745	4745	4745
R-squared	0.287	0.288	0.288

Note: Standard errors are in brackets; the significance level are *** p < 0.01, ** p < 0.05, * p < 0.1, one-tailed test.

show that the main effect of CSR is positive (α = 6.448, p < 0.01), and the quadratic effect of CSR is negative (α = −8.220, p < 0.01), supporting H1. The founder CEO interacts not only with CSR (α = −5.707, p < 0.10) but also with its quadric term (α = 10.704, p < 0.05), supporting H2.

5.4.3. Lagged Model

Considering that there is a possibility that CSR and firm performance cause each other, we lagged market values one period (MV_t+1) as dependent variable and tested our hypothesis again. Empirically, it is impossible for this year’s market values to cause CSR in the last year. This way, we can reduce the concern of mutual causation.

Table 8. Results with lagged MV as the dependent variable.

Variables	Model 1	Model 2	Model 3
	MVt+1	MVt+1	MVt+1
CSR	9.552 ***	9.439 ***	12.558 ***
	(2.223)	(2.224)	(2.755)
CSR2	−11.728 ***	−11.494 ***	−16.386 ***
	(3.332)	(3.334)	(4.261)
FCEO		0.480 *	1.471 **
		(0.274)	(0.629)
CSR × FCEO			−7.802 *
			(4.076)
CSR2 × FCEO			11.859 *
			(6.275)
Board size	−0.039	−0.026	−0.035
	(0.129)	(0.130)	(0.130)
Independent directors	−0.043	−0.066	−0.086
	(0.426)	(0.426)	(0.426)
Asset-liability ratio	−0.693	−0.685	−0.696
	(0.936)	(0.935)	(0.935)
Enterprise size	0.118	0.112	0.126
	(0.230)	(0.230)	(0.230)
Enterprise age	−0.207	−0.200	−0.126
	(1.241)	(1.241)	(1.241)
Enterprise growth	0.899 ***	0.902 ***	0.897 ***
	(0.217)	(0.217)	(0.217)
Degree of ownership concentration	−1.555	−1.648	−1.766
	(1.920)	(1.920)	(1.924)
Industry FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes
Constant	39.581 ***	39.409 ***	39.149 ***
	(4.510)	(4.510)	(4.511)
Observations	4014	4014	4014
R-squared	0.302	0.303	0.304

Note: Standard errors are in brackets; the significance level are *** p < 0.01, ** p < 0.05, * p < 0.1, one-tailed test.

shows that the main effect of CSR is positive (α = 9.552, p < 0.01), and the quadratic effect of CSR is negative (α = −11.728, p < 0.01), supporting H1. The founder CEO interacts not only with CSR (α = −7.802, p < 0.10) but also with its quadric term (α = 11.859, p < 0.10), supporting H2. That test indicates that our hypotheses are all valid and robust.

5.4.4. Endogeneity Test

To further address the endogeneity concern, we conducted several additional tests. First, we conduct two-stage least squares (2SLS) regression to mitigate the endogeneity concern [87]. Based on the literature [83], we use the industry average of the company in the current year (Ind_CSR) as an instrumental variable. As

Table 9. Results with instrumental variable.

Variables	Model 1	Model 1	Model 1
	First Stage	First Stage	Second Stage
	CSR t	CSR2t	MVt
Ind_CSR	0.883 ***	−0.090
	(0.227)	(0.146)
Ind_CSR2	0.082	1.239 ***
	(0.436)	(0.281)
CSR			27.853 ***
			(10.622)
CSR2			−35.187 **
			(17.559)
Board size	0.004 ***	0.003 ***	0.046
	(0.001)	(0.001)	(0.065)
Independent directors	−0.004	−0.006 **	−0.411 *
	(0.005)	(0.003)	(0.242)
Asset-liability ratio	−0.179 ***	−0.065 ***	1.521
	(0.009)	(0.006)	(1.011)
Enterprise size	0.010 ***	0.005 ***	0.376 ***
	(0.002)	(0.001)	(0.094)
Enterprise age	−0.001	−0.003	−0.379 *
	(0.004)	(0.003)	(0.216)
Enterprise growth	0.046 ***	0.014 ***	−0.065
	(0.004)	(0.003)	(0.330)
Degree of ownership concentration	0.044 ***	−0.001	−0.876
	(0.012)	(0.007)	(0.741)
Industry FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes
Constant	−0.044	−0.003	17.719 ***
	(0.033)	(0.021)	(1.297)
Observations	4745	4745	4745
R-squared	0.214	0.134	0.202

Note: Standard errors are in brackets; the significance level are *** p < 0.01, ** p < 0.05, * p < 0.1, one-tailed test.

shows, the instrumental variables in the first stage are valid. In the second stage model, the main effect of CSR is positive (α = 27.853, p < 0.01) and the quadratic effect of CSR is negative (α = −35.187, p < 0.05), confirming the inverted-U shaped impact of CSR on firm performance.

Second, we exchange the positions of the independent variables and dependent variables to further exclude the reverse causality between CSR and corporate performance.

Table 10. Results with interchange dependent variables and independent variables.

Variables	Model 1
	CSRt
MV	0.0004
	(0.0003)
Board size	0.0028
	(0.0020)
Independent directors	0.0068
	(0.0066)
Asset-liability ratio	−0.1307 ***
	(0.0138)
Enterprise size	0.0148 ***
	(0.0036)
Enterprise age	−0.0131
	(0.0178)
Enterprise growth	0.0466 ***
	(0.0035)
Degree of ownership concentration	0.1042 ***
	(0.0294)
Industry FE	Yes
Year FE	Yes
Firm FE	Yes
Constant	0.0425
	(0.0701)
Observations	4745
R-squared	0.213

Note: Standard errors are in brackets; the significance level are *** p < 0.01, one-tailed test.

shows that when the independent variable is MV and the dependent variable is CSR, the MV coefficient is 0.0004 and is not significant. It confirms that firm performance cannot influence CSR, thus reducing the endogeneity problem between CSR and corporate performance.

Third, to address selection bias, we conducted the two-stage Heckman test. Specifically, we lagged market values one period (MV_t+1) as the dependent variable and split the sample based on the mean CSR into two groups. The group assignment was predicted in the first stage model to calculate the inverse Mill’s ratio (IMR). The IMR was then entered into the second stage model.

Table 11. Results with Heckman test.

Variables	First Stage	Second Stage
	NCSRt	MVt+1
CSR		15.462 **
		(7.865)
CSR2		−17.118 *
		(9.164)
Board size	0.061 ***	0.150
	(0.020)	(0.132)
Independent directors	0.005	−0.829 **
	(0.073)	(0.337)
Asset-liability ratio	−2.468 ***	−2.590
	(0.147)	(3.893)
Enterprise size	0.068	−0.247
	(0.065)	(0.318)
Enterprise age	0.058 **	0.383 **
	(0.028)	(0.153)
Enterprise growth	0.379 ***	0.854
	(0.059)	(0.627)
Degree of ownership concentration	0.796 ***	0.951
	(0.174)	(1.416)
Mills lambda		1.431
		(2.623)
Industry	Yes	Yes
Year	Yes	Yes
Constant	18.080 ***	18.080 ***
	(6.599)	(6.599)
Observations	4419	4419

Note: Standard errors are in brackets; the significance level are *** p < 0.01, ** p < 0.05, * p < 0.1, one-tailed test.

shows that after considering the IMR coefficient, the second-stage test results show the main effect of CSR is positive (α = 15.462, p < 0.05) and the quadratic effect of CSR is negative (α = −17.118, p < 0.10), the inverted U-shaped relationship between CSR and corporate performance is further verified.

5.5. Mechanism Analysis: Financial Constraints as a Channel

We further explored a possible channel, financing constraints, through which CSR affects firm performance. First, regarding the relationship between CSR and financing constraints, it can be understood from two aspects: (1) Based on the stakeholder theory, CSR is relevant for many stakeholders, as it can improve the relationship between firms and stakeholders, help firms obtain more financing channels [88], reduce the information asymmetry between firms and stakeholders, attract more external investments [89], and alleviate the situation of corporate financing constraints. (2) Based on agency theory, the essence of corporate governance is to coordinate the rights and interests of corporate stakeholders. The process of corporate fulfillment of CSR will consider stakeholder demands [90], and corporate investment decisions will be more prudent and focus on long-term CSR investments to improve corporate investment efficiency [91]. As a result, the problem of corporate financing constraints can be alleviated.

Second, the relationship between financing constraints and financial performance can also be understood from two aspects: (1) In terms of capital structure [92], financing constraints limit the financing ability of firms and increase the difficulty and cost of capital structure adjustment, which is not conducive to the improvement of production and operation and financial performance. (2) In terms of capital liquidity, financing constraints put burdens on firms’ liquidity management and increase firms’ demand for liquid assets. When a firm has low liquid assets, financing constraints lead to greater financial risk and reduce the firm’s return on investment in liquid assets [93].

Based on the above rationale, it can be speculated that CSR activities are conducive to the improvement of the financing environment, enhancing the financing ability of enterprises, thus reducing financing constraints. At the same time, financing constraints negatively affect financial performance, and intermediary mechanisms are formed in the process of CSR to enhance financial performance. That is, financing constraints are a channel through which CSR affects financial performance.

We use the WW index to measure financing constraints [92] and use stepwise regression to verify the role of financing constraints as a channel through which CSR affects financial performance. As column 1 of

Table 12. Results of the mechanism analysis.

Variables	Model 1	Model 2	Model 3
	MVt	WWt	MVt
CSR	5.292 ***	−0.298 ***	6.703 ***
	(1.809)	(0.010)	(2.000)
CSR2	−7.271 ***	0.341 ***	−8.884 ***
	(2.808)	(0.016)	(2.972)
WW			4.733 *
			(2.860)
Board size	0.005	−0.001	0.009
	(0.110)	(0.001)	(0.110)
Independent directors	−0.057	0.004 *	−0.074
	(0.357)	(0.002)	(0.357)
Asset-liability ratio	1.129	0.021 ***	1.031
	(0.769)	(0.004)	(0.771)
Enterprise size	0.176	−0.026 ***	0.297
	(0.194)	(0.001)	(0.208)
Enterprise age	−0.014	0.011 **	−0.065
	(0.971)	(0.005)	(0.972)
Enterprise growth	0.841 ***	−0.044 ***	1.049 ***
	(0.199)	(0.001)	(0.236)
Degree of ownership concentration	2.852 *	0.108 ***	2.342
	(1.605)	(0.009)	(1.634)
Industry FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes
Constant	36.344 ***	−0.809 ***	40.171 ***
	(3.770)	(0.021)	(4.422)
Observations	4745	4745	4745
R-squared	0.285	0.665	0.286

Note: Standard errors are in brackets; the significance level are *** p < 0.01, ** p < 0.05, * p < 0.1, one-tailed test.

shows, the main effect of CSR is positive (α = 5.292, p < 0.01), and the quadratic effect of CSR is negative (α = −7.271, p < 0.01), verifying the inverted-U shaped relationship between CSR and firm performance. As column 2 of Table 12 shows, the main effect of CSR is negative (α = −0.298, p < 0.01), and the quadratic effect of CSR is positive (α = 0.341, p < 0.01), verifying the U-shaped relationship between CSR and financing constraints. As column 3 of Table 12 shows, the main effect of CSR is positive (α = 6.703, p < 0.01), the quadratic effect of CSR is negative (α = −8.884, p < 0.01), and the WW is positive (α = 4.733, p < 0.10). Put together, these results demonstrate the partial mediation effect of financing constraints on the relationship between CSR and financial performance. This suggests that CSR influences financial performance partially through financing constraints.

6. Discussion

By analyzing the 10-year data of a sample of China’s GEM listed companies, we discover an inverted U-shaped relationship between CSR and firm performance. As CSR increases, firm performance first increases until it reaches the highest value and then decreases. Our analysis further shows that the curvilinear relationship only exists for firms with non-founder CEOs and CSR has a linear relationship with firm performance for firms with founder CEOs. In addition, we find that financing constraints play a intermediary role between CSR and firm performance, suggesting that financing constraints serve as a channel through which CSR affects firm performance. These findings have important implications for not only research but also practice.

6.1. Implications for Research

This study makes two contributions to the research of entrepreneurial enterprises. First, it expands the knowledge of linear relationship between CSR and firm performance and enriches the research on the economic consequences of CSR by analyzing the nonlinear impact of CSR on firm performance. We find that there is an inverted U-shaped relationship between CSR and firm performance in China’s GEM companies. These findings confirm that there is an optimal input level for CSR which makes firm performance reach the maximum level [28]. Before the turning point, CSR increases performance, whereas after the turning point, CSR reduces performance. Thus, our findings confirm the complexity of the relationship between CSR and firm performance. Considering the background of encouraging universal entrepreneurship in China, focusing on China’s GEM companies, this research analyzes the impact of CSR on firm performance and broadens the scope of the CSR research. Based on stakeholder theory and resource dependence theory, this research analyzes the inverted U-shaped relationship between CSR and firm performance and further confirms that financing constraints are mediating variables in the process of CSR affecting firm performance. Our approach on the one hand provides a basis for promoting multi-theoretical analysis in the field of CSR, and on the other hand, provides a reference for exploring the mechanism of CSR’s influence.

Second, we discover a boundary condition for the inverted-U shaped relationship by showing the moderating effect of founder CEO. Previous research on the impact of CSR on firm performance produced mixed results [20,24,27,30,94], showing that the impact can be linear, non-linear, or non-significant. Our findings could shed light on the ambiguous relationship between CSR and firm performance, which help to reconcile the inconsistencies. In firms led by founder CEOs, CSR has a linear impact on firm performance, indicating that the presence of founder CEO ensures a consistent response of firm performance to CSR. We posit that founder CEOs’ stewardship role may explain why this happens. Under founder CEOs, although the increase of firm performance is slower in response to increased CSR activities, there is no decrease of firm performance. In contrast, under non-founder CEOs, although the increase of firm performance is faster in response to increased CSR activities, there is a turning point after which firm performance quickly declines as CSR keeps increasing. These findings deepen the understanding of the role of founder vs. non-founder CEOs in CSR management. From a corporate governance perspective, our study is different from previous studies that focus on personal characteristics, such as CEO age, tenure, and integrity or basic board structures, such as board size, the proportion of independent directors, and the proportion of female directors. The purpose of our study is to understand how CEOs’ moderating effects depends on whether they are founders or not. Our findings deepen the understanding of the role of CEOs in the context of CSR, expand the scope of CSR research by integrating research on leadership and corporate governance, and creating new theoretical insights by drawing from agency theory and stewardship theory.

6.2. Implications for Practice

Under current institutional environments around the world, more and more enterprises are actively undertaking CSR. This research has implications for entrepreneurial practices, particularly in CSR investment activities. First, it shows that CSR can indeed lead to improvement of firm performance for China’s GEM listed companies. However, excessive investment in CSR is a stumbling block to improved firm performance. Enterprises playing the role of “citizen” in the market and undertaking CSR should prevent excessive investment in CSR, so as to avoid the negative impact on firm performance. Most important, the start-ups should strive to find the optimal amount of CSR input so that they can fully absorb its positive effects to improve firm performance, while in the meantime avoid overinvestments on CSR that hurt their performance.

Second, it also suggests that enterprises should recognize the differences caused by founder vs. non-founder CEOs. Compared with firms under non-founder CEOs, firms under founder CEOs can consistently improve their firm performance by increasing CSR investment. While non-founder-CEOs can strengthen CSR’s effect in improving firm performance before the turning point, it could lead the firms to a pitfall and make them falsely believe that further performance improvement can be gained by increasing CSR investment. However, it turns out that after the turning point they would suffer a performance dip with more CSR investments. Based on what types of CEO they have, firms can design different strategies to pursue maximal firm performance. If they have a non-founder CEO, they should be alert to whether they have reached the turning point and take a more conservative CSR stance after the turning point. If they have a founder CEO, they can be more aggressive in CSR investment given that the founder CEO will play a stewardship role to oversee CSR investments.

In addition, we find that in our sample the firms with a founder CEO tend to be younger and smaller, while firms with a non-founder CEO generally are older and bigger. This suggests that the founder CEO seems to have a greater role in the start-up period of a firm and that along with the gradual growth of the firm founder CEO may be replaced by non-founder CEO. Thus, our findings may ring alarms to incumbent firms, which are often led by professional CEOs because these firms are vulnerable to overinvestments on CSR. As they focus too much on CSR, their financial performance could be impaired.

6.3. Limitations and Future Research

This research has some limitations, which can provide ideas for future research. First, as far as the sample is concerned, we only examine China’s GEM listed enterprises. Future research should expand the sample size, focus on the global market, and combine the data of start-ups in more countries to verify the economic consequences of CSR. Second, for the moderating effects, we only examined the presence of founder CEO. Founders may also have other management roles, e.g., founder chairman of the board of directors. CEOs also have different characteristics, such as gender and education level. It is unclear whether the effect of CSR on firm performance is contingent on these factors. We encourage more future research to explore in depth the impact of founders holding other management positions and CEO characteristics on the relationship between CSR and firm performance. Third, the measure of CSR in our research is a composite indicator without distinguishing CSR types. Future research is needed to examine in detail the financial and economic effects of different types of CSR [95]. A meta-analysis can be carried out first to identify which types of CSR are most significantly related to firm performance. Fourth, we only examined financing constraints as a mechanism through which CSR affects corporate financial performance, and there are other possible channels to be explored, such as corporate reputation. Future research can combine the reputation mechanism to explore more channels through which CSR affects financial performance.

7. Conclusions

Based on the analysis of 10-year panel data of start-up public companies in China, following the stakeholder theory and resource dependence theory, this research analyzes the nonlinear relationship between CSR and firm performance. Further, we combine agency theory and stewardship theory to excavate the moderating role of the founder CEO in the relationship. In addition, we test the mediating role of financing constraints in the relationship between CSR and firm performance. There are three main findings of our study: (1) CSR input has an inverted-U shaped impact on firm performance. There is an optimal amount of CSR input that can maximize firm performance. (2) The nonlinear impact of CSR on firm performance is only present in firms with non-founder CEOs, while the impact of CSR on firm performance is linear for firms with founder CEOs. (3) Financing constraints play a partial mediating in the relationship between CSR and firm performance, suggesting that CSR can improve firm performance because it alleviates the financing constraints of companies and subsequently improve firm performance. This paper provides empirical evidence on how CSR affects firm performance and how this relationship is moderated by the presence of founder CEO. The findings have great importance to management theory and the practice of entrepreneurial corporate governance in light of the current social environment that vigorously encourages enterprises to undertake CSR.