Corporate Social Responsibility in Hotels: A Proposal of a Measurement of its Performance through Marketing Variables

: The research into Corporate Social Responsibility (CSR) has been proliﬁc in the last years, although few studies have focused their attention on studying its relationship with economic performance within the hotel industry, even less incorporating marketing variables as a result. This work aims to determine the relationship between the implementation of CSR policies and their inﬂuence on the Revenue Per Available Room (RevPAR) in Spanish hotels, through marketing variables. A quantitative methodology was conducted using an online survey, gathering a sample of 230 hotel managers from Spain. The results of the structural model analyzed reveal that CSR has a discreet but signiﬁcant role for understanding how marketing variables and RevPAR operate. CSR has a direct impact on RevPAR, but it also inﬂuences it indirectly through marketing variables. Therefore, CSR arises as a fundamental strategy to improve the results of the hotel sector in the long term.


Introduction
Corporate Social Responsibility (CSR) focuses on the benefits for businesses and society alike. All the definitions of the concept point out that firms must satisfy both the stockholders' and the stakeholders' interests [1]. Although CSR has recently received greater attention and has been studied from different approaches, its application to the tourism and hospitality sector and specifically to the hotel industry is still relatively scarce [1][2][3]. Even though most of the research are focused on the relationship between CSR and performance [4], its results have not been conclusive, possibly because it is a multidimensional concept that changes according to the context in which it is applied [5].
The measurement of the impact of CSR in hotels has been made mainly from point of view of ROA (Return on Asset)-the short-term result-or the Tobin's q-the long-term result-based on secondary data obtained from databases [6,7]. Although these studies have been applied to hotel companies, they have not taken into account the main variable of the financial performance of this sector, the Revenue Per Available Room (RevPAR), which measures the financial performance of an establishment or chain for a given period (RevPAR = total income per room in period t / total number of rooms available in period t). The studies on CSR have been traditionally developed through a three-dimensional construct, which includes an economic, environmental and social facets (e.g., [2,[8][9][10][11][12]), although they differ in some cases in the allocation of these dimensions. This paper analyzes the impact of the three dimensions of CSR on hotels' performance, measured by RevPAR. In addition, marketing variables are also involved in the model with the aim of assessing their influence on results variables. This work falls into the research stream identified by Serra-Cantallops et al. [1], including this study into the category "CSR impacts from consumer (Marketing) and company perspectives (Firm Business-Performance)".
The objective of this study is to determine the relationship between the use of CSR's policies and the performance studied through RevPAR in Spanish hotels, through marketing variables such as reputation, trust, satisfaction and loyalty, from the perspective of the manager. In short, it is a model that aims at analyzing the influence of CSR in the hotel sector, integrating marketing and financial results.

Trust
Trust is meant to be a requirement for the development of long-term relationships between companies and clients [13], and a guarantee that the hotel will act in a competent and reliable manner [14].
Martínez and Rodríguez del Bosque [14] proposed a model where CSR influences loyalty, through trust and satisfaction as mediators, among other variables, showing the direct and indirect effect between these constructs. Lombart and Louis [11] investigated the impact that CSR has on retail establishments and the price image on other variables such as satisfaction, trust and intentions of future behaviors, among other variables, getting to test the influence of CSR on trust. In the context of the banking sector, Fatma et al. [15] also demonstrated the effect of CSR on trust.
Hypothesis 1 (H1). CSR has effects on the trust in the hotel.

Reputation
Reputation has been confused with image or considered as synonymous with it. Reputation is defined as a "mental scheme that synthesizes, organizes and simplifies the clues offered by the multiple images projected by the company" [8] (p. 43).
There is a general agreement that CSR activities can directly translate into reputation for the organization [8,16,17] and that it can be an important source of competitive advantage [8]. Zhu et al. [17] researched antecedents and results of CSR with a sample of hotels and travel agencies in China. The results showed that CSR is positively related to the reputation of the company. Their results are in line with those achieved by Kim and Kim [3], Melo and Garrido-Morgado [18] and Su et al. [19].
Hypothesis 2 (H2). CSR has effects on the reputation of the hotel.

Satisfaction
Probably the most accepted definition in the literature on satisfaction is: "the consumer's response to a degree of compliance" [20] (p. 13). However, Bowen and Clarke [21] ensured the need to treat satisfaction according to the specificity of every context in which it may arise. Garay and Font [9] maintained that numerous academic studies defend that CSR has a positive correlation with financial performance as a competitive advantage in the accommodation sector, positively affecting reputation and consumer satisfaction, among others. In the same line, Kang et al. [6] ensured that CSR can improve customer satisfaction and loyalty. Lee and Heo [22] affirmed that CSR has a positive impact on customer satisfaction, but also on the value of the company. between 25% and 85%, depending on the sector, retaining 5% of its clients. In the hospitality sector, there is a great interest in identifying which factors determine loyalty to the hotel, because it is generally estimated that loyal customers lead to greater profitability [14].

Hypothesis 10 (H10). Loyalty has effects on RevPAR.
In their study Lee and Heo [22] examined the mediating effect of satisfaction between CSR activities and performance, confirming that for hotels CSR activities have a positive impact on the satisfaction and value of the company.

Hypothesis 11 (H11). Satisfaction has effects on RevPAR.
Inoue and Lee [7] tested the relationship between certain dimensions of CSR and financial performance in tourism companies. Their results revealed that each dimension has a different effect on both the short-and long-term benefits, and that the financial impacts vary according to the four types of companies studied (airlines, casinos, hotels and restaurants). While according to the results achieved by Zhu et al. [17] the reputation of the company is positively related to performance.

Hypothesis 12 (H12). Reputation has effects on RevPAR.
Kang et al. [6] pointed out that there is a significant relationship between CSR and the performance of the company in the hospitality industry. However, Zhu et al. [17] stated that the inconsistent results achieved so far on the effect of CSR on the performance of the company point to the need of deepening into the research in this field.
García and Armas [27] analysed the importance of social and environmental responsibility in the company and its relationship with performance, because the authors argued that the literature offers contradictory results, especially in the hotel sector, given the intense relationship that links this industry with the environment. Their results suggested a strong and positive relationship between both variables. Hypothesis 13 (H13). CSR has effects on RevPAR. Figure 1 shows the theoretical model that has been described. In their study Lee and Heo [22] examined the mediating effect of satisfaction between CSR activities and performance, confirming that for hotels CSR activities have a positive impact on the satisfaction and value of the company.

H11: Satisfaction has effects on RevPAR.
Inoue and Lee [7] tested the relationship between certain dimensions of CSR and financial performance in tourism companies. Their results revealed that each dimension has a different effect on both the short-and long-term benefits, and that the financial impacts vary according to the four types of companies studied (airlines, casinos, hotels and restaurants). While according to the results achieved by Zhu et al. [17] the reputation of the company is positively related to performance.

H12: Reputation has effects on RevPAR.
Kang et al. [6] pointed out that there is a significant relationship between CSR and the performance of the company in the hospitality industry. However, Zhu et al. [17] stated that the inconsistent results achieved so far on the effect of CSR on the performance of the company point to the need of deepening into the research in this field.
García and Armas [27] analysed the importance of social and environmental responsibility in the company and its relationship with performance, because the authors argued that the literature offers contradictory results, especially in the hotel sector, given the intense relationship that links this industry with the environment. Their results suggested a strong and positive relationship between both variables.
H13: CSR has effects on RevPAR. Figure 1 shows the theoretical model that has been described.

Scales' Selection
The indicators used to measure the model have been obtained from the literature review. The scale on CSR is based on the one proposed by Martínez et al. [12], and the trust concept on the one used by Martínez and Rodríguez del Bosque [14] and Verhoef et al. [28]. The scale of reputation was proposed by Alvarado and Schlesinger [8]. The scale of loyalty has been supported by the proposal of Martínez and Rodríguez del Bosque [14], with the introduction of a new indicator based on the literature review (LOY5) [29]. For the evaluation of performance, a scale has been proposed that combines the indicators proposed by Chavarría [30] (REV1 and REV2), Grissemann et al. [31] (REV3) and Garay and Font [9] (REV4). The measure of performance has been made through RevPAR. All indicators of the model are considered as reflective. For their measurement a 7-point Likert scale was used. The final indicators used and their codes are exposed in Table 1 in the fourth section.

Research Scenario
This study was located in Spain. According to the most recent available data of the World Tourism Organization (WTO) [32], Spain received 83 million arrivals in 2018, which makes this country the third tourist destination worldwide [32]. The total population of hotel establishments published in the Hotel Occupancy Survey elaborated by the National Institute of Statistics [33], according to data from February 2020 (last data available), is 12,662 hotel lodgings, offering a total of 1,185,245 beds, mainly placed in the category of four-star (554,251) and three-star hotels (261,201) [33]. With regard to the profitability of the national hotels the RevPAR is 49.43€. The inter-annual variation rate of RevPAR reveals that the sector is booming if the annual average for 2019 is analyzed.

Fieldwork and Data Analysis
In this study, we have opted for a quantitative methodology using an online survey. A total of eight academics participated in the pretest of the questionnaire. A piloting of the same was also performed, with a subsample of six hotel directors. The dissemination of the questionnaire was carried out by e-mailing 8,410 Spanish hotels, addressed their managers, with telephone support to encourage a higher response rate. Through a non-probabilistic convenience sampling, 519 surveys were started, but only 230 were considered valid responses for having been completed. That makes a response rate of 2.7%. The fieldwork was carried out from 8th of April to 1st of September 2015.
Since this work has an exploratory approach, and because of the introduction of CSR as a second-order construct with three dimensions, the use of Partial Least Squares (PLS) seems to be the best option. The sample size obtained, composed by managers of Spanish hotels, is acceptable according to the parameters required by PLS [34]. The software used was SmartPLS 2.0.
A sample of Spanish hotels was used, offering a good representativeness of their territory. Of the hotels involved in the research, 33.9% are four stars, followed by three stars (29.1%). Regarding the type of management, the majority (101) are independent hotels, followed by family-run hotels (81) and hotels belonging to large chains or franchises (40). Table 1. Assessment of the measurement model.

Results
The proposed model presents CSR measured with three dimensions-economic, social and environmental-hence the model includes second-order constructs. Its assessment requires the application of the two-step approach [35,36]. Firstly, the model is evaluated taking into account that the first-order factors will act as the second-order construct they represent. After that, it is necessary to use the scores of the indicators of the dimensions of the CSR as new aggregate indicators of each dimension, to proceed with the second step, evaluating again the measurement model. In both steps, indicators were considered reflective.
With regard to the analysis of individual reliability, the indicators must have a load equal to or greater than 0.707 [34].
In the first step, all the indicators exceed this threshold, with an exception in ENV5 "uses renewable energy" (0.6189). Hair et al. [34] state that usually indicators with loadings between 0.40 and 0.70 should only be eliminated if this leads to the improvement of composite reliability (ρ c ). Given the improvement that occurs in the composite reliability in the ENV construct (from 0.9084 to 0.9107), the ENV5 indicator has been dropped. In the second step, the individual reliability is verified for all the items. With regard to composite reliability, these values should be between 0.60 and 0.70 in exploratory studies [37]. All of the constructs have very high values for the composite reliability, both in the first and in the second step. The convergent validity has been assessed by analysing the value of the Average Variance Extracted (AVE), which must be higher than 0.5 [34]. All the AVE values for each construct are greater than the limit value marked in both steps (see Table 1).
The evaluation of discriminant validity is done by demonstrating that the correlations between the constructs are lower than the square root of the AVE [38] (see Tables 2 and 3). To evaluate the structural model it is necessary to analyze the R 2 for each dependent construct, and the significance of the paths using a bootstrapping procedure [34]. The explained variance of an endogenous construct by another latent variable is also reported. This value is given by multiplying the path coefficient by the correlation coefficient between both variables [39]. Table 4 shows the values of R 2 . The model has a moderate-substantial explanatory capacity for the outcome variables (loyalty and performance) and a weak-moderate explanatory capacity for the intermediate variables (trust, reputation and satisfaction). Yes † R 2 value of 0.75, 0.5 or 0.25 for the latent endogenous variables in structural models can be considered substantial, moderate or weak, respectively [34]. ‡ Critical t-values: * p < 0.05; ** p < 0.01; *** p < 0.001; ns not significant (based on t(4999), two-tailed test); t(0.05;4999) = 1.65; t(0.01;4999) = 1.96; t(0.001;4999) = 2.58.
Regarding the hypotheses, the test of the proposed causal relationships give empirical support to the model, except for the paths linking CSR-satisfaction (H 4 ), CSR-loyalty (H 6 ), trust-loyalty (H 7 ) and satisfaction-RevPAR (H 11 ). These results award a discrete role to the CSR as a direct predictor of marketing variables, but it also demonstrates an impact on performance through RevPAR in hotel establishments. For this reason, it is understood that relationships that occur between the implementation of CSR and performance (RevPAR) and marketing variables cannot be dissociated. Figure 2 offers a graphical resume of model assessment.

Discussion
With regard to the relationship between CSR and trust (H1) (21.9%), it is proven to be statistically significant. This result is consistent with those found by Lombart and Louis [11] in retail establishments, Fatma et al. [15] in the banking sector, Melé et al. [40] in the case of the environmental dimension of CSR and Martínez and Rodríguez del Bosque [14] in the hotel sector. These latter authors emphasized that clients are more inclined to believe in socially responsible companies that operate honestly in their activities and care about the interests of both parties of the relationship in decision making. Therefore, the CSR becomes an effective tool to create trust between the client and the company.
Although CSR exerts a greater direct effect on trust (21.9%), rather than on reputation (12.1%), this last variable seems to have a more prominent role, especially because of the indirect effects produced by the CSR through reputation on satisfaction, loyalty and performance. The relation CSR-reputation (H2) turns out to be significant in this study, a result that is in line with those obtained by Zhu et al. [17]. However, this finding was not proven by Alvarado and Schlesinger [8], probably due to the fact that their study was carried out on a different activity (mobile telephone usage).
Trust has a strong weight in the formation of satisfaction (25.5%), but less than reputation (36.3%). The results related to the relationship trust-satisfaction (H3) are consistent with those achieved by Martínez and Rodríguez del Bosque [14].
The relationship CSR-satisfaction (H4) is not supported in the context of this study, unlike what happens in the studies developed by Martínez and Rodríguez del Bosque [14] and Lombart and Louis [11]. This could be due to the fact that the studies by these authors employ samples composed

Discussion
With regard to the relationship between CSR and trust (H 1 ) (21.9%), it is proven to be statistically significant. This result is consistent with those found by Lombart and Louis [11] in retail establishments, Fatma et al. [15] in the banking sector, Melé et al. [40] in the case of the environmental dimension of CSR and Martínez and Rodríguez del Bosque [14] in the hotel sector. These latter authors emphasized that clients are more inclined to believe in socially responsible companies that operate honestly in their activities and care about the interests of both parties of the relationship in decision making. Therefore, the CSR becomes an effective tool to create trust between the client and the company.
Although CSR exerts a greater direct effect on trust (21.9%), rather than on reputation (12.1%), this last variable seems to have a more prominent role, especially because of the indirect effects produced by the CSR through reputation on satisfaction, loyalty and performance. The relation CSR-reputation (H 2 ) turns out to be significant in this study, a result that is in line with those obtained by Zhu et al. [17]. However, this finding was not proven by Alvarado and Schlesinger [8], probably due to the fact that their study was carried out on a different activity (mobile telephone usage).
Trust has a strong weight in the formation of satisfaction (25.5%), but less than reputation (36.3%). The results related to the relationship trust-satisfaction (H 3 ) are consistent with those achieved by Martínez and Rodríguez del Bosque [14].
The relationship CSR-satisfaction (H 4 ) is not supported in the context of this study, unlike what happens in the studies developed by Martínez and Rodríguez del Bosque [14] and Lombart and Louis [11]. This could be due to the fact that the studies by these authors employ samples composed by customers, while the present study is focused on executives, who may have a more limited vision of the satisfaction of their customers, since it is an estimation.
As noted above, reputation has a significant effect on the formation of satisfaction (H 5 ) (36.3%), which is a result consistent with the findings of Zhu et al. [17].
The CSR-loyalty relationship (H 6 ) resulted to be not significant, which confirms the results obtained by Lombart and Louis [11]. However these results are not in accordance with those obtained by Kucukusta et al. [25], Martínez et al. [12] and Melé et al. [40], who proved that relationship. In this case, it is not possible to find a potential cause in the limited estimation of hotel managers can make of their customers' loyalty, since these studies used a sample of hotel customers. Therefore, it is necessary to continue deepening in this relationship in future studies.
The relationship between trust-loyalty (H 7 ) turned out to be not significant, a result that also occurred in the study by Lombart and Louis [11]. As well as in the present study, Palacios-Florencio et al. [41] concluded that when trust acts as a mediator between CSR and loyalty, the direct relationship between these two last variables turned out to be not significant. These results differ from those found in by Martínez and Rodríguez del Bosque [14]. They succeed in demonstrating that trust is a strong determinant of loyalty. Moreover Martínez et al. [42] found significant the relationship between trust and behavioral intentions in the context of the environmental dimension of CSR. Thus, they affirm the importance of environmental certifications as a strategic tool for fostering trust and positive customer intention behaviors in hotels. Meanwhile, in the context of destination social responsibility, Su et al. [43] found that tourists with a strong sense of trust have a greater intention to visit the destination, a finding that highlights the importance of trust a as a key driver that extends the application of organizational trust in the tourism sector.
The link found between satisfaction and loyalty (H 8 ) is particularly relevant, since satisfaction contributes by 60.0% to the formation of loyalty. The relationship between satisfaction and loyalty has been widely verified in the literature through multiple studies in the tourism sector [44]. It is noteworthy that this result is consistent with the outcomes achieved in other models involving CSR, as is the case of the study by Lombart and Louis [11], which also proves a positive relationship between satisfaction and behavioral intentions and loyalty.
Once again, reputation corroborates its relevant role in the relationship with loyalty (H 9 ). Even if its impact is quite weak (9.7%), if compared with that exerted by satisfaction (9.7%), this relationship is significant, compared to the non-significance found for the relations between CSR-trust and CSR-loyalty. The results of the reputation-loyalty relationship are consistent with the findings of Zhu et al. [17]. Su et al. [43] also provided valuable insights about the importance of reputation in the context of destination social responsibility, showing that when a destination has a good reputation it offers a positive effect on tourists' trust and intentions to visit the destination.
The results of this study support the relationship between loyalty and RevPAR (H 10 ), with loyalty being the main antecedent of performance measured through RevPAR (17.9%). These results are in line with those reached by Reichheld and Sasser [26] and Martínez and Rodríguez del Bosque [14]. This result could underscore the importance of investing in loyalty programs to boost performance. To this regard, Hua et al. [45] tested how loyalty programs influence hotel performance. Their findings show that the investment in loyalty programs has a significant and positive impact on RevPAR, among other performance indicators.
Despite the relevance of satisfaction for marketing theory, and although it has an important role in determining loyalty, in the present model, it has a non-significant impact on another variable of the model, such as RevPAR. This confirms that the hypothesized causal path involving CSR-satisfaction-RevPAR (H 4 and H 11 ) is not significant, that is, satisfaction does not exert a mediating effect between CSR and performance. The results obtained for this sequence are in agreement with those obtained by Lee and Heo [22] who explored the same causal sequence.
Reputation is confirmed as a relevant variable in the model, as the relationship reputation-RevPAR (H 12 ) (16.2%) is proven to be significant and the variable is the second strongest antecedent of RevPAR, only preceded by loyalty, that boasts a slightly higher percentage (17.9%). This result is consistent with the work of Zhu et al. [17] as well as the fact that CSR exerts an indirect impact on performance through reputation. That is, the CSR-reputation-performance causal path is confirmed both in the work of these authors and in the present study.
Finally, the last hypothesis, linking CSR and RevPAR (H 13 ), is proven to be significant (7.2%) in the context of this research, in line with what was previously confirmed by Kang et al. [6] and García and Armas [27]. However, its effect is weaker than the one exerted by loyalty or reputation on RevPAR.
This model tries to offer new empirical insights from the marketing perspective to improve the hotel performance and highlighting how marketing variables such as trust, reputation and satisfaction can constitute a sustainable strategy over time, achieving loyal customers who promote economic performance and ensuring the continuity of the company.

Conclusions
This paper presents a model for the study of CSR that combines marketing variables and economic results that the CSR implementation generates for the company. RevPAR is used as a measure of performance, as it is a commonly accepted indicator in the sector. On the other hand, the measurement of CSR through the economic, social and environmental dimensions is applied, following the operationalization proposed by Martínez et al. [12]. Its application to a sample of hotel managers, selected in the present research, gave back positive results. The results of this study are in accordance with the results obtained by García-Pozo et al. [46]. They concluded that "customers of hotel services have an increasing interest in business attitudes that promote the greater involvement of hotel establishments in their environmental, social, and economic setting" (p. 10).
On the other hand, the scale used for the measurement of RevPAR, as an indicator of performance in the hotel sector, also resulted to be appropriate. The scales used for trust, satisfaction and loyalty, based on Martínez and Rodríguez del Bosque [14], and reputation, based on Alvarado and Schlesinger [8], showed good results in their adaptation to the study context, so they can be taken into account for further investigations. Regarding the results of the model's assessment, CSR, even if discrete, has a significant role in the understanding of the functioning of some variables.
CSR has been proven to be a significant antecedent for trust and reputation, and a determining factor for RevPAR. However, although it is understood that the impact is discrete, the relationships between CSR and performance (RevPAR) could not be disregarded, nor could its relationship with traditional marketing outputs.
The results obtained in the present study have important practical implications and inspire some recommendations for the management of hospitality companies. First of all, it is proven that CSR policies have a positive influence on RevPAR and on long-term performance. The sector is encouraged to develop this type of activity, in its economic, social and environmental dimensions. Hotel companies can develop CSR by initiating energy efficiency programs, management and separation of waste at source, control of spills, purchasing policies with sustainable criteria, promotion of local gastronomy and culture, continuous training of employees, policies of gender equality, labor climate surveys etc.
It is important to consider that CSR is an activity that should be promoted by top management, and that it should reach all departments of the organization, and that in turn not only has repercussions on customers, but on all the stakeholders of the company.
The limitations of this study refer to the generalizability of results, which are limited by the use of a non-probabilistic and convenience sampling method. In relation to the sample, it is worth noting that the work focused on the hotel managers' perspective. This implied that, with regard to some variables, respondents had to answer the questionnaire according to an estimation of the demand perspective. Regarding the model, there are possibly other significant variables related to CSR, such as quality and image, that have not been taken into account in this research.
This study could be replicated in other contexts, sectors and activities within the tourism sector, as well as in other countries. It could also be interesting to incorporate the consumers' perspective to the proposed model. Future works can go in the line of progressing in the knowledge of CSR, in order to provide the sector with effective tools capable of enhancing its performance and its competitive permanence in the market over the long term.