Does a Company’s Proﬁtability Inﬂuence the Level of CSR Development?

: The objective of this paper is to analyze the effect of economic and ﬁnancial performance on Corporate Social Responsibility (CSR). For this reason, we have used the data from a sample made up of 662 companies, 146 registered as medium-sized or large and 516 as small or micro, highlighting the signiﬁcant weight of small companies in the sample. CSR has been measured using an indicator estimated from the data gathered by way of a questionnaire containing information related with the economic, environmental, and social dimensions. The analysis has been conducted by estimating panel regression models with robust errors. The results show a negative relationship between economic performance and more CSR activities implemented, supporting the Managerial Opportunism Hypothesis. Furthermore, large companies under the pressure of stakeholders are more prone to implementing certain CSR actions than small ones, meaning that a minimum size is essential according to this research.


Introduction
There are many available definitions of Corporate Social Responsibility (CSR), and often no clear definition is given, making theoretical development and measurement difficult; more specifically, Dahlsrud [1] analyzes 37 definitions. From the start, this could already make us think that it is a problem. However, if they are analyzed in more detail, as this author indicates, the definitions are predominantly congruent. It also points out that the different CSR definitions describe a phenomenon, but do not present any guide on how to manage the challenges of said phenomenon, and the definitions are context independent. According to Van Marrewijk [2], a successful CSR strategy must be specific to the context of each individual business.
Corporate Social Responsibility (CSR) is defined in the renewed EU Strategy 2011-2014 for Corporate Social Responsibility as "the responsibility of companies for their impact on society" (COM/2011/0681 final). The strategy clearly establishes that "respecting the legislation applicable and collective agreements between social partners is a prerequisite for complying with this responsibility and focuses its content on the integration of social, environmental and ethical concerns, respect for human rights and consumers' concerns in their business operations and their core strategy, in order a) to maximize the creation of shared value for their owners/shareholders and for other stakeholders and society at large; and b) to identify, prevent and mitigate its possible adverse consequences". The Organisation for Economic Co-operation and Development (OECD) has introduced an alternative term, Responsible Business Conduct (RBC), which is defined as "making a positive contribution to economic, environmental and social progress with a view to achieving sustainable development and avoiding and addressing adverse impacts related to direct and indirect operations, products or services of a company" (OECD, 2011). In this way, CSR or RBC mean the responsibility in the management of organizations considering social, environmental, and labor problems, in such a way that economic development and society are favored by it.
There is increasing awareness of CSR among stakeholders in organizations: Customers, employees, and society. Thus, MORI's CSR Research [3] found that 70% of European consumers believed that a company's commitment to social responsibility was important to them when buying a product, and that 20% would be willing to pay more for sociallyand environmentally-responsible products. On the employee side, companies with high CSR have greater capacity to attract and retain qualified human resources. In the scientific literature, a positive relationship was shown for job satisfaction and CSR in companies [4] or [5], in workers' performance [6], and in job search intentions [7]. Finally, companies that meet CSR criteria reduce the risks derived from socially or environmentally irresponsible behavior (fines, financial penalties, etc.), while the use of CSR practices indicates better management quality. Thus, there are several reasons that lead companies to implement CSR: To respond to stakeholders' specific demands, to improve company performance and corporate reputation, to generate customer loyalty, and to prevent legal sanctions [8]. CSR activities have been recognized in recent years as a natural obligation of firms [9].
The relationship between CSR and the financial performance of companies has been extensively addressed in the literature. Hategan, Sirghi, Curea-Pitorac, and Hategan [10] summarize the research about the relationship between financial performance and CSR. However, most of the studies on CSR have focused on large companies, with a few studies considering micro, small, and medium-sized enterprises (see, e.g., Hernández and Sánchez [11], and Socoliuc et al. [12]). Micro and small enterprises also carry out CSR activities, but more studies on the nature and dynamics of CSR in smaller enterprises need to be undertaken [13]. Furthermore, most of the studies have analyzed CSR as an explanatory variable. In order to explain the level of CSR implementation, delayed performance has been used in this work. In this regard, one of our aims is to contrast the hypotheses of managerial opportunism and availability of funds, to see to what extent the most profitable companies are committed to CSR. In addition, our research suggests that large companies have greater human and financial resources that allow them to implement a greater number of plans, while small enterprises may have more difficulties. Besides this, another of our aims is to analyze the effect of the financial performance on the three dimensions of CSR (social, labor, and environmental).
With this in mind, this work contributes to the literature in several aspects. Firstly, it provides evidence of whether financial performance determines the level of CSR development in the company. Secondly, the influence of company size on this relationship is analyzed, an aspect which has often been overlooked in the research conducted to date, their having focused on analyzing large companies. Thirdly, CSR indices are estimated for companies whose data is not easily accessible, since they do not report sustainability information. This paper is structured as follows: Section 2 describes the theoretical framework and the research hypotheses; in Section 3, the methodology used is explained in detail, describing the sample, the independent and dependent variables, and the models analyzed; the results are shown and discussed in Section 4. Finally, the conclusions of the study are discussed.

Theoretical Framework and Hypotheses
The relationship between CSR and performance has been studied in depth in the economic literature, with various hypotheses being proposed. As we will explain below, there are different theories, on the one hand, it analyzes how CSR influences performance, on the other, there are proposals that analyze the effect of performance on CSR, and finally, studies that analyze a bidirectional effect between both variables.
Regarding the possible effects of CSR on performance, there are theories that postulate a positive relationship, that is, companies most involved in CSR should have higher financial performance. The Stakeholder Theory (see Freeman [14]) is included in these arguments, which suggests that companies that are more CSR-oriented will please the different stakeholders more and, therefore, will achieve a better reputation, and in turn will result in greater performance. Similarly, the reputation of companies that do not meet their stakeholders' needs will deteriorate, and consequently, their performance will worsen too [15,16]. This theory is the basis of the approach of good management [16] or of social impact [17] according to which, CSR improves relationships with the main interest groups, generating a best overall result. There are also conflicting approaches which maintain that the relationship should be negative. This group includes the approach of Friedman [18], under the name of "trade-off", that considers the costs associated with CSR's place in the company in a less competitive position, reducing its profitability. Ultimately, what underlies this theory is that the costs associated with these activities exceed the possible additional income that it could generate.
In turn, the causal relationship is also questioned when it is considered that it is the company's performance that determines the CSR level. On this subject, the hypothesis of the Availability of Funds indicates that the relationship that can be found between both factors is positive, but that profitability ultimately determines CSR spending. In this case, financial performance is usually considered an exogenous and lagged variable.
McGuire et al. [19] concluded that financial performance is a variable that influences acts of social responsibility and, more specifically, that the previous financial performance of companies is closely related to corporate social responsibility. A similar result was obtained by Roberts [20], whereby the approach of this hypothesis stated that the better the result a company obtains, the higher the margin will be, and that they will be able to invest in any of the dimensions covered by CSR [16,21]. In short, when carrying out some type of social activity, it will depend in part on the desire and social commitment of the company, but also on whether there are resources available for it. This implies that low profitability can limit the company's capability to allocate resources to CSR activities. Following Ullmann [22], periods of low profitability mean that the different economic demands of a company may take priority over the expenses associated with activities of social responsibility. However, the Managerial Opportunism Hypothesis poses the opposite relationship. It is relatively common that the objectives of the owners and other stakeholders may be different from the objectives of the managers, often with the latter focused more on the short term and the more immediate result. This theory argues that when managers earn a lot with the company, they try to maximize their private profit by spending little on CSR, which is against the interests of other stakeholders [23]. Under this approach, companies with high economic performance could reduce their involvement in CSR activities to maximize their personal income [24]. The company could increase its profitability by reducing social spending, while other, unprofitable companies might become more involved in CSR to justify their poor results.
In the existing literature, the possibility of a bidirectional relationship between both variables has also been considered. Moreover, if the Stakeholder Theory and the Availability of Funds are simultaneously given, a virtuous circle between both variables would be produced, causing a bidirectional and positive effect. This relationship, known as The Positive Synergy Hypothesis, is contrasted in the work of Waddocd and Graves [16]. The conclusion reached in the work of Martínez-Ferrero and Frías-Aceituno [25] is similar, evidencing the existence of a synergistic circle. On the other hand, if the hypotheses of the Trade-off and the Managerial Opportunism are combined, a virtuous circle would be produced between both variables, causing a bidirectional effect, but in this case negative.
In general, the results of the studies undertaken give out mixed, albeit positive and significant, signals in terms of the relationship between CSR and corporate financial performance (CFP), although there are divergences in the results. Pava and Krausz [26], Orlitzky et al. [27], Wu [28], Gössling [29], and Wang et al. [30] showed that there were mixed results in the relationship between CSR and Economic Performance (EP), namely positive, negative, or neutral. Based on the evidence of 21 empirical studies, Pava and Krausz [26] noted that twelve of them had found a positive association, eight had found no association, and only one had found a negative association. Similarly, Gössling [29] reviewed multiple studies, noticing that there was a positive relationship between CSR and CFP in 67% of the cases, there was no significant relationship in 29% of the studies, and a negative and significant relationship was seen in only one of the studies analyzed (4%). Among the studies that conclude the bidirectionality between the variables, we highlight Orlitzky et al. [27], and in the recent work by Busch and Friede [31], 25 meta-analyses were examined with a total number of 1902 primary studies, where a positive, strong, and bidirectional relationship between CSR and CFP was witnessed.
Taking into account all the above, a great variety of the conclusions reached in the different studies has been deduced, occasionally finding a positive, negative, or nonsignificant correlation, and with different causal directions.
These differences in the literature have been attributed to variations in the CSR measurement (such as the use of different indicators or dimensions) and in the CFP measurement (in accounting or market performance), as well as the moderating factors that affect the relationship between CSR and CFP, similar conclusions are provided in the paper of Wang, Dou, and Jia [32].
For his part, Marom [33] concludes that the relationship between both variables is not linear, and that the inverted U model would explain the relationship. As this author indicates, such a non-linear model can explain the range of different outcomes. This relationship could be positive or negative, depending on the different levels of corporate CSR.
In this work we have focused on analyzing the effect of financial performance in CSR. In this sense, we try to contrast the Availability of Funds and Managerial Opportunism Hypothesis. Furthermore, traditionally papers in this field have focused more, and therefore are more numerous, on studying the effect of the implementation and development of CSR on other variables, among which is performance. In short, they are more focused on how CSR can add value to the firm. In our case, the objective is another. We focus on how the characteristics of the company, and more specifically financial performance, can affect CSR.
Therefore, we propose that: Hypothesis 1 (H1). The effect of a company's profitability on the CSR level is positive.
The increasing interest in CSR is down to the different stakeholders of an organization asking to receive reports on its behavior in order to reassure themselves that it is transparent and ethical. At a business level, there are several methodologies for evaluating the dimensions of sustainability; however, the most accepted at an international level is that provided by the Global Reporting Initiative (GRI). This independent non-governmental organization created the first world standard for drafting sustainability reports, which helps organizations understand and communicate their impact to their stakeholders. Drafting the sustainability report and its publication on the GRI website is voluntary. After consulting this database, it has been observed that in Spain at the time of writing there were 280 large companies (1549 reports), 56 medium-sized enterprises (227 reports), and 270 micro and small enterprises (808 reports) that had drafted and communicated their GRI report. Since most of the studies have used the so-called GRI reports, there is a bias in the research towards the CSR of larger entities, as these types of organizations are under greater sociopolitical pressure or are keen to demonstrate their CSR. Smaller enterprises generally lack CSR knowledge [34] or do not feel a need to publish reports on it. Thus, they are analyzed in very few studies [30]. Herrera, Larrán, and Martínez [35] carried out a meta-analysis, learning that of the 258 resulting articles after their screening, only 18 related the CSR of SMEs with their performance. Among these examples of research, Niehm et al. [36] found a significant relationship between size and the community support dimension which, in turn, maintained a significant and positive association with the gross annual income. Marín and Rubio [37] analyzed the effect that each CSR dimension has on competitive success in companies with between 5 and 250 employees. In this case, the control variable size was not significant, so if we consider only small enterprises for instance, size does not act as an intensifier of the positive influence that CSR has over competitive success. The other side of the coin is that in a subsequent study [38] in which SMEs and large companies were considered, the size factor did turn out to intensify the positive influence that CSR had on competitiveness. Similarly, papers such as those of Godfrey et al. [39], Sharma [40], Aguinis and Glavas [41], and Brotons and Sansalvador [42], among others, have indicated that as size increases, the relationship between CSR and performance is strengthened, which they call the moderating variable. These last authors [42] also conclude that CSR is not exclusive to large businesses, but firm size determines the size of the impact of the CSR on the results.
Worthy of note is that most of the studies that have analyzed CSR determinants do consider, among other variables, size. On this subject, in the investigations of Bansal [43], Castelo and Lima [44], Reverte [45], Chih et al. [46], Pozniak et al. [47], Wang et al. [48], and Cebolla et al. [49], the size variable was found to be significant in the adoption of CSR measures. These results can be justified based on the Resource Availability Hypothesis, with regard to both material and human resources, which can have an influence, firstly, over the ability to host CSR activities and secondly, over publicizing them. In addition, other reasons that would justify the positive effect of size on social responsibility according to Graafland, Van de Ven, and Stoffele [50] would be that large firms are more visible to the public and the media, therefore, responsible behaviors will have a greater impact on large companies. Furthermore, as a company grows, it generally means that it becomes more visible, and consequently, has more responsibility to meet the demands of the stakeholders. Likewise, the existence of greater economies of scale in large companies would cause the costs related to the development of CSR to be lower for these companies. Additionally, the larger ones will have a greater need to make relatively more use of formal and public instruments to communicate its responsibility, while smaller companies would use more informal channels, or even as relations with stakeholders are more personal and direct, they do not communicate the different actions related to CSR. In this sense, and according to the Theory of Silent Corporate Social Responsibility [51], smaller companies could have more actions related to CSR in place than it might seem at first glance. Therefore, if only formal channels and official reports are taken into account, it is very likely that the relationship is significant and positive, but if a search is made of all the activities related to CSR, both formal and informal, it could be that the variable size was no longer significant.
Legitimacy theory also explain the levels of CSR participation. Following this theory, external stakeholders require the enterprise to take CSR actions that are desirable and are in alignment with society's values (see Suchman [52]). The more visible the companies are, they need to conduct more CSR initiatives to increase their legitimacy. Following this theory, Udayasankar [53] proposed U-shape relationship between the levels of CSR participation and the firm size, i.e., high levels of CSR participation of small and large firms and lower levels for middle-sized companies. Large firms are highly visible, so a low CSR participation would bring unfavorable impact, however, small-sized firms are less visible firms, but they are involved in CSR for gaining legitimacy.
On the other hand, in highly competitive environments, the assumption of new costs derived from CSR will generally have greater repercussions on smaller companies as their sensitivity is greater and would make them less competitive [54]. In addition, the less separation between ownership and control and the less professionalization of management in smaller companies would favor that the commitment to CSR depends more on the values and personal beliefs of the owner-manager in this type of company [51].
Taking into account all the above, we propose the following hypothesis: Hypothesis 2 (H2). The size of the company positively affects the level of CSR implementation.
The vast majority of authors agree that the implementation of different measures related to CSR requires using funds. Consequently, the availability of funds can be related to a higher level of liquidity and thus, if more liquid resources are available, it can help different activities to develop related to CSR. In this sense, the liquidity variable could help us analyze the hypothesis of the Availability of Funds. We consider this variable as a good indicator of the availability of real resources that could be invested immediately. It should also be pointed out that the availability of resources will not always materialize in greater liquidity, it could happen that the investment of those idle resources occurs more immediately and therefore this increase in liquidity does not occur. Or it could even be that those companies with greater liquidity are precisely due to the fact that they do not carry out investments in CSR, and that is the reason why they have more liquidity.
Authors such as Jennifer and Taylor [55], Abd-El Salam and Weetman [56], Wallace and Naser [57], Owusu-Ansah [58], Talha et al. [59], and Oyelere et al. [60], based on the signaling theory, have considered that companies with more liquid resources may have stronger incentives to provide more details about the corporate tasks they are carrying out. The results of the studies by Cooke [61] and Wachira [62] indicate that there is a positive and significant relationship between liquidity and information disclosure related to CSR. Nevertheless, the results obtained by Jennifer and Taylor [55] were the opposite of their initial predictions in that higher liquidity was significantly and negatively associated with CSR disclosure. On the other hand, in the papers of Kahl [63] and Talha et al. [59], the liquidity variable was not significant. Another reason for this situation to occur could be owing to the fewer liquid resources available being allocated more wisely and being invested in CSR-related activities among other areas, resulting in fewer idle resources. To illustrate this point, DiSegni el al. [64] concluded, contrary to their expectations, that the companies that made up the Dow Jones Sustainability Indices (DJSI) were characterized by statistically significant lower liquidity measures.
Based on the above, we propose the following hypothesis:

Hypothesis 3 (H3).
The liquidity of a company positively affects its CSR level.

Methodology
In order to find out the relationship between performance and CSR, we need to have an indicator that allows us to differentiate between companies that have the greatest involvement, not only on a global scale, but also in a wide range of CSR areas. Thus, this section explains the construction of a CSR index. We also describe the data used and the independent variables analyzed. Since the data available on the performance is prior to that relating to corporate responsibility, our study examines how a company's profitability affects the level of CSR development. Thus, the aim is to analyze whether the most profitable companies foster CSR to the largest extent (the Hypothesis of the Availability of Funds) or whether they cut their spending on it, in line with the proposals of the Managerial Opportunism Hypothesis. This relationship is analyzed with a sample that contains companies of various sizes.

Data
The study has been performed out of a sample of Spanish companies participating in the Thirteenth Corporate Social Responsibility Report of the Xunta de Galicia's permanent CSR observatory (see Table 1). The information sources come from a questionnaire given by email and over the telephone at the end of 2018. This questionnaire collected data related to the implementation of CSR and the companies' relationship with the environment, as well as the annual accounts, from where the economic and financial indicators were obtained. As a consequence, primary (CSR survey) and secondary (SABI, Iberian Balance Analysis System) sources of information have been combined. To compile the questionnaire, different CSR reporting standards were primarily taken into account, mainly GRI Indicators. In order to analyze the size of the companies, they were classified according to the Recommendation 2003/361/EC of the European Commission.

Dependent Variable: CSR Index
We have used dichotomous variables as representative variables of the level of Corporate Social Responsibility (Table 2) implementation and have built an index that measures the development level of good CSR practices. Therefore, we have developed our variable for CSR via the partial indicators of each dimension. For the social sphere, eight variables have been measured, twelve for the labor and eight for the environment. Therefore, a company that has implemented all the aforementioned aspects will score a maximum of 28 points. The CSR indicator (CSRIND) has been built up from the sum of three subfactors that include aspects related to conflict resolution, social action, equality, conciliation, insertion, environmental indicators, and certifications.

Dimension Dichotomous Variables Analyzed
Social Customer service. Consumer arbitration service.
Claims sheet. Promotes CSR among its customers. Encourages responsible consumption among its customers.

Identifies and evaluates impacts of its activity on the territory
Has an anti-corruption policy. Develops actions in favor of social development.

Labor
Presence of equality plans. Activities to promote equality for groups with difficulties.
Presence of training plan. Promotion based on previously-defined criteria.
Wage policy with perks. Reduced working hours. Flextime. Working hours matching children's school schedule.
Teleworking. Job sharing. Extension of paternity/maternity leave. Job placements for the disabled, immigrants, young people, etc. Therefore, it is positively valued when a company has conflict resolution mechanisms, promotes CSR practices among customers, favors social development, evaluates its impact, encourages work-life balance, and operates in an inclusive manner. Using environmental indicators, following corrective measures, holding environmental certifications, and not having sanctions is also considered.

Independent Variables: Performance Measures
Studies such as Moskowitz [65] and Rennings, Schröder, and Ziegler [66] used the market price variation of companies to evaluate the relationship between performance and CSR. The use of such a measure excluded unlisted companies from the analysis, so performance measures based on accounting and not based on the market have been considered. Thus, measures such as financial profitability (ROE) and economic profitability (ROA) are the most frequently used indicators in the literature [67,68], although other authors have added new ratios, and even Moore [69] used an accounting measure with an index composed of several ratios. In our work, we have used various ratios like economic performance measures, including financial profitability (ROE) and economic profitability (ROA). The table below (Table 3) shows the definition used for each variable, with SABI being the data source for all of them. In turn, other variables such as size, solvency level, and liquidity in which the company operates have been considered. Table 3. Variables used to measure performance.

Variables Definition
Financial profitability or ROE Operating income/Equity Economic profitability or ROA EBIT/Total assets Size Logarithm of total assets Liquidity Current assets/Current liabilities Solvency Equity/Non-current assets Source: Own elaboration.

Descriptive Statistics
The descriptive statistics of the variables used in this work are shown in Table 4. The mean value of the CSR indicator is 13.45, although there are significant differences between the companies based on size. Large companies have more than 19 points, while small enterprises only have a score of 12. In general, these differences are the same for the three index components, more noticeable for labor. The same occurs with profitability, but in this case, medium-sized enterprises show the best performance, both for ROA and ROE, with large and micro-SMEs having the worst performance.  Table 5 shows the relationship between the different variables used in this work, measured with Spearman's Correlation Coefficient. In general, a positive correlation is observed, although it is more intense between "labor" and "social", and far less so with environmental issues. This implies that greater sensitivity with social aspects in a company is consistent with involvement in the workplace, to a lesser extent with environmental aspects, although the relationship is positive all the same. Regarding the relationship with performance, the coefficients are close to zero, although it should be taken into account that this relationship may be affected by other variables. To evaluate the differences in the level of CSR implementation based on profitability, a dummy variable has been estimated based on whether the company's ROA exceeds the median or not. Subsequently, we have performed a test for the mean differences for the total sample, differentiating by size ( Table 6). As can be seen, the most profitable companies have significantly lower values in the general CSR index and in the environmental one, showing a negative relationship between performance and CSR.

Estimated Models
The analysis has been conducted by estimating panel regression models with robust errors, considering the year and the company, and by using accounting information related to the period 2014-2017 to evaluate the effect of different delays. The data relating to the CSR level corresponds with the year 2018. The estimated model is formed as follows: where: CSR i = this measures the development level of CSR in company i at moment t.
Performance it−1 = ROA or ROE of company i at moment t − 1. X it = variables related to size, debt, and liquidity. ε it = Error term.

Results
The results (Table 7) obtained in our study show a negative relationship between the profitability obtained by the company and the level of involvement in CSR, and thus supports the Managerial Opportunism Hypothesis. To recap, this proposes that managers put their private objectives before those of the remaining stakeholders [23], and is in line with the findings of Preston and O'Bannon [17], Cespa and Cestone [70], and Martínez-Campillo, Cabeza-García, and Marbella-Sánchez [24]. Furthermore, these results contradict the Hypothesis of the Availability of Funds, which postulates the positive influence of profitability on CSR. Other variables have proven to be significant and with a positive sign, such as the size of the company and liquidity. Size shows that larger companies increase the level of CSR activity, obtaining higher values in the index. This could be due to the need to have a minimum size to implement certain operations that are not available to the smallest enterprises. In addition, larger companies have greater union and regulatory pressure, which in many cases forces them to become more involved at social, labor, and environmental levels [43][44][45][46][47][48][49]. Regarding liquidity, the results obtained are in line with the papers of Cooke [61] and Wachira [62], which may well indicate that more liquidity companies have, the more resources they have at their disposal to invest in CSR activities.
When we analyze the CSR areas, we observe that the negative effect in the labor is significant for large companies, but not for small ones.

Discussion
Traditionally, the literature has analyzed the role of CSR in companies' performance. The mixed evidence found in the literature between CSR and firm performance may be a result of inconsistency in defining CSR, inconsistency in defining firm performance, inconsistency in samples, and changes over time, among others [71]. In addition, it is commonly accepted (e.g., Lantos [72]) that there are three types of CSR, known as "ethical CSR", "altruistic CSR", and "strategic CSR". CSR minimizes firms' harm (ethical CSR) and promotes benefits for society (altruistic or philanthropic CSR if the firm does not reciprocally benefit, and strategic CSR if it does). Hillman and Keim [73] make a distinction between altruistic CSR and strategic CSR, finding a positive relation between firm performance and strategic CSR and a negative relation between altruistic CSR and firm performance. Our work focuses on the relationship between the profitability obtained by the company and the level of involvement in CSR. Our previous results suggest a negative relationship between economic performance and CSR. Hence, from our results, the conclusion that companies with higher economic and financial performance implements more CSR activities and that less profitable companies reducing them cannot be drawn. The relationship found could be linked to the trade-off hypothesis put forward by Friedman [18], and that managers might choose not to invest because the costs associated with CSR would place their company in a less competitive position and reduce its profitability. If profitable companies reduce their level of involvement in CSR, this may imply that they consider it unprofitable to carry out CSR since companies end up incurring higher costs that are difficult to recuperate. This relationship is observed both globally and in the samples that divide large and small enterprises. Our results are consistent with the fact that several companies think that enterprises are responsible only to their shareholders, and not to society in general. We have not been able to distinguish between the different types of CSR, and future studies seem necessary to analyze the relation between economic performance and different types of CSR.
When we analyze the CSR areas, we also observe a negative effect in the labor, environmental, and social dimensions. However, if we consider the effect of company size, labor is significant for large companies, but not for small ones. Therefore, the Managerial Opportunism Hypothesis is fulfilled when the company is larger, which implies a reduction in resources and actions aimed at improving the working conditions of its employees. This, however, does not occur with micro enterprises, where the level of CSR development does not depend on profitability. This result could be owing to many micro-SMEs having just started their operations and may show a greater commitment to CSR, flexible working arrangements and work-life balance for example, regardless of the company's results. The relationship with the environmental index is also not significant, which could be indicative of the greater commitment that companies are making with the environment regardless of the resources they are capable of generating.
Furthermore, large companies under the pressure of stakeholders are more prone to implementing CSR actions than small ones. This may be also due to corporate social reporting, as only large companies make these statements. In our case, when we use a survey to collect data on CSR, we understand that we collect data from both formal and informal CSR actions. Therefore, our results would not be biased by the possible existence of the silent social responsibility theory. Laws and regulations to encourage companies to submit more CSR information in their financial statements and annual reports would boost CSR actions implementation. Companies can get more out of non-altruistic CSR activities by publicizing them, using CSR as a marketing device (a form of public relations). Our study also supports the inverted U shape relationship with CSR participation and size of Udayasankar [53].
Regarding liquidity, the results indicate that the more liquidity companies have, the more resources they have at their disposal to invest in CSR activities. This behavior would be justified by the hypothesis of the availability of resources. The different entities consider the availability of more available liquid resources as a key factor in carrying out CSR initiatives.

Conclusions
In this paper, we have evaluated the effect of delayed financial performance as an explanatory variable of the level of CSR implementation. It differs from previous studies on this subject in that most of them have focused on large companies, and that CSR has been analyzed as an explanatory variable. The results obtained show a negative relationship between the profitability obtained by the company and the level of CSR implementation (measured as an index that measures the development level of good CSR practices), the opposite to that obtained by Hategan et al. [10] for listed companies. These results support the Managerial Opportunism Hypothesis. The above relationship is observed both globally and in the samples that are split into large and small enterprises. The relationship found could be linked to the trade-off hypothesis put forward by Friedman [18], and to managers choosing not to invest because they believe that the costs associated with CSR make the company less competitive and lose profits.
This could mean that companies analyze CSR investment projects from the perspective of financial value creation and not from a cost-benefit analysis (see Bosch-Badia, Montllor-Serrats, Tarrazon-Rodon [74]). On the other hand, the relationship between CSR, company size, and liquidity are significant and positive. Regarding size, all signs point to the likelihood of minimum size being essential to implement certain initiatives that small businesses cannot assume and, furthermore, large ones are subject to greater pressure due to being more in the public eye. The same relationship is seen for liquidity, which means that companies with greater liquidity have a more comfortable level of resources, which they can allocate to activities related to CSR.
When the different CSR areas are analyzed, the effect in labor is only negative and significant for large companies. This behavior could be justified with the Managerial Opportunism Hypothesis, in that when the company is large, it reduces resources allocated to labor. The relationship is not significant for the environmental dimension, which could be read as a greater commitment from companies regardless of their performance.
Measuring the development level of good CSR practices as an index, and not the level of investments in CSR actions as a percentage of profits or expenses, may be considered as a limitation of the paper. Another limitation is not considering the level of R&D spending and advertising expenditure, since these variables have been shown to be determinants of firm performance (e.g., McWilliams and Siegel [75]).