A A a study about the social actions in the bRazilian Retail sectoR

The purpose of this article is to analyze the social actions undertaken by Brazilian retail companies. The study is justified by the need to have an on-going assessment of actions of this nature, bearing in mind that over the last 15 years many retail companies have been developing many social initiatives. The empirical data were taken from the database of information recorded by the Centro de Excelência no Varejo, of the Fundação Getulio Vargas. In order to assess the data, an adaptation of the James Austin (2000) model was made. Analysis of the data was predomi-nantly exploratory, and involved the descriptive analysis, factor analysis, and cluster analysis. It was also made the cross analysis of the clusters generated with some of the variables. The results indicated a strong relationship between the social and economic value generated by the actions of companies and characteristics relating to their size, partnerships and the geographic spread of the actions.


abStr act
The purpose of this article is to analyze the social actions undertaken by Brazilian retail companies. The study is justified by the need to have an on-going assessment of actions of this nature, bearing in mind that over the last 15 years many retail companies have been developing many social initiatives. The empirical data were taken from the database of information recorded by the Centro de Excelência no Varejo, of the Fundação Getulio Vargas. In order to assess the data, an adaptation of the James Austin (2000) model was made. Analysis of the data was predominantly exploratory, and involved the descriptive analysis, factor analysis, and cluster analysis. It was also made the cross analysis of the clusters generated with some of the variables. The results indicated a strong relationship between the social and economic value generated by the actions of companies and characteristics relating to their size, partnerships and the geographic spread of the actions.

KEy wordS
Social and environmental responsibility; Social and environmental action; Retail companies; Economic value; Social and environmental value.

iNtrodUctioN
The attention that companies in Brazil pay to social issues is growing. In fact, as one of their priorities many companies choose to meet the demands of the less privileged members of the public, by developing specific programs for these communities. In this context many retail companies develop social initiatives to encourage the economic and social development of the communities they have chosen. The Getulio Vargas Foundation has been recording these efforts and they have been analyzed by the Centro de Excelência no Varejo (GVcev). Since 2003 this center has been sponsoring the Prize for Social Responsibility in the Retail Sector, thereby recognizing the role that retail companies play in this area vis-à-vis Brazilian society.
This Prize recognizes and encourages the social responsibility programs or actions developed by retailers throughout Brazil. These actions reveal the commitment and the ethical position the retailer adopts in relation to its activities and to society. The inclusion of social responsibility in the management of these businesses involves strengthening the relationships of the retailer with its various interested publics, be they employees, consumers, suppliers, stockholders, partners, government, the community and even the environment, with the aim of guaranteeing the sustainability of the company over the long term.
The GVcev organized the initiatives of those retailers that had been submitted for the prize award to the Database of Social Responsibility Best Practice in the Retail Sector. This database offers a research tool that has various search resources that are available for consultation by those who are interested. The cases submitted are presented in summary form and highlight some of the exemplary social action practices coming out of the Brazilian retail sector.
In this work we present an analysis of the initiatives on this database that were submitted for the prize award. The research used a protocol prepared from proposals made by Carroll (1979), Wood (1991), Schwartz and Carroll (2003) and Austin et al. (2004), who analyzed aspects of corporate social performance. At the beginning we present in summary form the conceptual basis we used. Then we set out the methodology and operational procedures used for obtaining and analyzing the data, the results we obtained and finally the recommendations that emerge as a result of the research.

tHEorEtical rEfErENcE PoiNt
The conceptual point of reference is organized into three main blocks: a) factors that encourage the company to act socially, b) evaluation of the results, and c) collaborative actions. Carroll (1979) initially proposed and subsequently elaborated in detail (CAR-ROLL, 1991) that a company's social responsibility should be evaluated in three distinct dimensions: categories of responsibility, the philosophy of responsibility and the social aspects involved.

the Factors that encourage responsible action
As far as the categories of responsibility are concerned the author proposed the existence of four categories: economic, legal, ethical and philanthropic. According to him these dimensions neither are mutually exclusive, cumulative, nor additive, as can be seen from the Figure 1. categories oF social resPonsiBility, according to carroll

Discretionary or philanthropic
Voluntary contributions made by the company with the aim of improving community life. They result from the company's desire to contribute and not because of any laws or from an ethical sense.
Ethical activities the company is involved in that go beyond their legal obligations. These responsibilities are difficult to assess and are the result of principles and standards that define the acceptable conduct as determined by the public the company is interested in. Carroll (1998) subsequently dealt with the relationships between the categories, pointing out that they are interrelated and frequently conflicts. Furthermore, according to him, the social responsibility of a company should be assessed using all of these categories.
Carroll identified another component of a company's social responsibility; the philosophy of responsibility. This aspect involves the philosophy, manner of doing things or strategy that leads the company to act in a responsible way. This philosophy may operate within a continuum that begins from a non-response situation, where nothing is done, to a position at the other extreme of pro-activity, where the company does something that is socially responsible.
In second place, according to Carroll, the company carefully chooses a social aspect where it should act after having decided in favor of responsible action. From the analysis of work done on this topic by Wilson (1975), McAdam (1973 and Davis and Blomstrom (1966) Carroll identified -and was supported in this by the results of Wilson (1975) -the idea that companies may act in accordance with one of four possibilities: reaction, defense, accommodation or pro-action. Figure 2 gives a synthesis of these possibilities. the PhilosoPhy oF resPonsiBility, according to carroll WILSON, 1975

Nothing is done < continuum > A lot is done
Source: Carroll (1979) In this way Carroll proposed that a company decides to act (or not) in a responsible way motivated at some level by the four basic situations of the continuum. It can decide that it should do nothing or if it decides to do something it can act in a reactive, defensive, accommodated or pro-active way.
Thirdly, as Carroll highlighted, various social issues are implicit in the responsible action of a company. These issues must be considered in any analysis of its actions. The above mentioned author, supported by the results obtained by Holmes (1976), warned that the social actions carried out by responsible companies may be associated with the five factors presented below: a) it combines a social necessity with a corporate ability for helping; b) it involves a relevant social need; c) it arises from the interest of the company's executives; d) it represents an opportunity or value for Public Relations; and e) it is associated with government pressure.

eValuating results
Evaluation of the results is an essential activity in any action, including the social responsible actions developed by a company (SANTOS, 2004). Initially it is essential to identify whether the company makes any type of effort to evaluate the results produced by its actions. Furthermore, as pointed out by Santos (2004), the evaluation may be looked at from different angles. It can be appreciated as to its social impact or be expressed in economic terms. This is the case, for example, if the social action generates some type of revenue.
On the other hand, within the sphere of the initiative and the impact it causes are aspects that are of interest to both the beneficiaries as well as to the organization that is carrying out the action. These aspects can be jointly considered. Finally, according to Santos (2004), the value generated by the action may interest one or more of the stakeholders with which the company has a relationship. Its actions may generate value for company employees, or for specific segments or stakeholders that are of interest to the company. Company suppliers or supporters may be involved in the actions. In addition, the actions may arise the interest of company allies, the community in which it operates or society as a whole.

collaboratiVe actions
As Austin (2000) and Austin et al. (2004) pointed out the responsible actions of a company may be carried out in a collaborative way, involving other companies, with civil society organizations and other agents of society. For example, the company may carry out actions in partnership with NGOs that operate in the community in which it has an interest, may operate with local government authorities and even use subsidies provided by these sources. The company may also work in partnership with other companies.
This tool was successfully used by Austin (2000) in his study of collaborative actions carried out in the USA. He also used it in partnership with researchers from various Latin American countries (AUSTIN et al., 2004). His proposal is that collaborative initiatives may be evaluated in accordance with the nature of the relationship and characterized in stages. Figure 3 identifies these dimensions and stages. According to the dimensions indicated above the stage of collaboration in which the company develops its actions may be characterized as follows: • Philanthropic stage. Austin points out that this is the most common situation in which companies find themselves. Putting it simply, the author explains that companies typically donate funds to social organizations in response to requests that the latter make to the company. Therefore, the nature of the relationship is reduced in all the dimensions considered. This is a demonstration of the company's concern with society at the lowest level. • Transactional stage. The author points out that in this stage we find the companies that are beginning to pay greater attention to social causes. This is when collaborative relationships become more intense and are of interest to both sides. Companies no longer merely pass their funds over; they start to concern themselves with establishing greater links between their social responsibility actions and their own activities. The author clarifies that in companies that are at this stage we find examples of social actions that are associated with the company's own marketing actions, the marketing of causes. • Integrative stage. A reduced number of companies are to be found in this stage, although as the author indicates, the tendency is for this area to grow. Actions assume the character of strategic alliances, missions that identify with each other, synchronized strategies and the sharing of values between the parties. Organizations start interacting with increasing frequency and intensity and make joint efforts. In this stage new joint efforts with different types of collaborative format are often created.
Austin alerted us to the fact that this scheme for assessing collaborative actions has been used to analyze the actions of responsible companies and he emphasized that the situation of a company may change according to the dimension analyzed. In other words, in a particular dimension the company may be in a particular situation and in others it may be in an advanced position. Therefore, the scheme is not rigid and the collaboration may present characteristics that are different depending on the dimension being considered.
These authors pointed out that the scheme has proved to be useful in situations in which various collaborative actions are being analyzed. This is because this scheme allows us to compare the patterns and identify any similarities between them. What matters for professionals and researchers, according to Austin and his colleagues, is to systematically identify and analyze the characteristics of the alliances so that decisions may be taken in the search for the generation of greater benefits from the social initiatives undertaken by companies, jointly with other organizations.

FranCIsCO JOsÉ Da COsTa • rUbEns Da COsTa sanTOs
In short, the results of the theoretical point of reference presented in this part allowed for the development of the protocol used in the research. This basis gave rise to the approach used to assess the commitment of retail companies to social responsibility. From it, and without requiring any further research with companies and those responsible for the actions that were contained in the database, we extracted simple indicators, whose presence could be evaluated in those cases submitted for the prize award and from the data available in the database.

data aNd MEtHodoloGy
We analyzed 160 cases from the two editions of the prize. When the companies registered they were already supplied the following data for identification purposes: company (assumed name), corporate name, category, retail sector, CNPJ (ID of the firm), contact name, position, telephone, e-mail, website, address, region, and program/project.
To collect the data on the cases we developed a research protocol, constructed from the theoretical reference point previously presented. The proposed form was analyzed by different research assistants and then submitted to a pre-test using 10 cases. The definitive script used in the research is shown in Figure 4. The details are set out below:  • In relation to the number of projects, as a measure we used the number of projects indicated by the companies. For those cases in which different and disperse actions were mentioned we considered all of them as being a single project. • As far as motivation is concerned, we looked for each of the indications in each case. If we detected any of them, it was shown as "yes". Therefore , there were cases in which all the indications of motivation were shown as being present. • Regarding the impact of the initiative, from the description of each case we tried to discover if the benefits had a strong, medium or weak impact as far as both the beneficiaries and the organization carrying out the social action were concerned. • With regard to the geographic level, we looked to see if the actions of the company had an impact at the local (community and urban region), regional (state to geographic region), or national (in more than one geographic region to the whole country) level.

figuRe 4 (continuation)
worKing variaBles • As for the value generated, we sought to identify both the social as well as the economic value. Both indications were classified as either, high, medium or low. • For indications of the collaboration existing in the action, we tried to check if the project was an initiative of the organization itself, if it had become involved in a project belonging to someone else, if there was collaboration with a non-governmental organizations, or with the government, or if it was carried out with other partners that were different from these last two.
Finally, with regard to the level of collaboration, we sought to check if the companies behaved in a philanthropic, transactional or integrative way (AUS-TIN, 2000;AUSTIN et al., 2004). So, by using a scale from 1 to 5 we looked to check for the presence of 8 indications. For the classification using these criteria we used the average number of indications for each criterion, with those companies that had averages of 1.66 or less being considered as philanthropic, those with averages between 1.67 and 3.33, as transactional, and those with averages over 3.34, as integrative.
For each of these elements we extracted, first, the basic descriptive statistics and then the crossings that were considered relevant. The companies we researched into were identified as retail bodies, (including bodies that were the representatives of other companies), large, medium, small and micro companies and shopping centers. The distributions of the large, medium and small companies were very close, so much so that the number of retail bodies was small.
With regard to the origin of the projects, we saw that half came from the State of São Paulo, with Minas Gerais (nearly 8% of the companies) in second place, followed by Paraná (7.5%). The other states did not exceed 5%, and some of the states in Brazil presented no projects at all.
With regard to the number of projects we saw that the companies indicated between 1 and 7 projects. We analyzed a total of 275 projects, with an average of 1.75 projects per company analyzed. As is shown in the following table the vast majority only presented 1 project and only one company presented 7 projects.

MotiVation
With regard to an evaluation of what motivates companies to develop social responsability actions, we saw that approximately three in every four companies sought to combine a social need with their corporate capacity for helping. This is an observation that is in line with what we expected from the companies and their projects, since it is more convenient for the company to develop actions that are in accordance with its competencies. It is relevant to observe, on the other hand, that one in every four companies we analyzed was identified as having an involvement with actions that are not directly associated with its corporate capacity for helping, as is the case of companies such as Açougue Cultural T-Bone, that, in spite of operating in the meat sales area, has developed a project associated with community libraries.
In analyzing the motivation associated with the level of social need towards which the projects were directed, we saw that nearly two thirds of the projects sought this orientation and were more associated with delicate social issues of the present day, like hunger, the care of patients with cancer and aids, support for street children and various others.
Also on this topic, we saw that one in every three of the companies we analyzed was not apparently driven by this particular motivation and was involved with projects that are also important (like sporting projects for customers, leisure for employees, etc), but that were not actually associated with social issues, like the ones mentioned in the previous paragraph.
Subsequently we checked if the involvement of companies with the projects was associated with issues that were of particular interest to the executives and saw that in 77% of the companies we looked at the actions were in accordance with interests such as obtaining a better reputation in the market, attracting customers and maintaining a relationship with them, improving the surrounding area where the company is located, etc.
With regard to this item, 23% of the companies we analyzed were seen to be developing projects and actions that were not in accordance with anything that was of interest to their executives. In other words, it seems that they were involved with actions in activities such as support for the body combating the use of drugs, donations to churches and other similar bodies and joint actions with them; these were in no way associated with the interests of their executives.
Still considering the orientation with regard to the category item we could see that the greatest number of actions associated with the interests of the executives occurred in retail bodies, followed by shopping centers and by medium size companies. In the other categories the percentages were close to the general percentage.
In the next item we tried to see if there was a more specific manifestation of interest on the part of the executives in order to check whether the actions in fact represented a Public Relations opportunity or value for the company. The reason for this analysis came from the perception that it was possible that the project was of interest to the executives, but did not represent the aforementioned opportunity or value. As we saw from the data we collected there is a degree of interest on the part of executives that is not necessarily apparent in the initiatives, albeit in few cases. In fact we saw a divergence in the 24 cases we analyzed. When we checked the association between the category of companies and the motivation in the Public Relations opportunity or value it was clear that the supermarket category represents the greatest percentage as far as not developing projects with this end in mind is concerned. This indicates that this category of companies develops more projects that are of interest to their executives, but that do not have the potential to be converted into public relations opportunities. These are projects directed at the categories as a whole, with a greater focus on affiliated or represented units, and not exactly at the body itself. On the other side of the coin, in other words, where the projects are more directed at the development or exploitation of public relations activities, we have Shopping Centers and medium size company categories.
Finally, we sought to check for the existence of any evidence that companies were motivated by government pressure. In this item we saw that only 24.4% of companies were motivated by this, in cases associated with projects that are of benefit to employees, the creation of access for physically handicapped people to the company's points of sale, or the recycling of waste.
In relation to this item we saw that the vast majority of companies develop their projects without receiving any government pressure whatsoever; the low percentages are completely the opposite of what we saw above, where the factor indicated was always one of the main motivations for the social actions. We can consider this aspect to be relevant, because it points to the fact that in their social responsibility actions the majority of companies seek to meet their own organizational and social objectives, and to a lesser extent meet the demands of government.

iMpact oF the initiatiVe
In the analysis of the level of impact of the initiatives, initially observing the impact on the beneficiaries, we considered that the majority of the projects had a weak impact, in other words, these were projects of relatively little importance and need, especially those that were more directed towards the development of the company's relationships with its customers. We could see that 27.5% of the cases analyzed have a large impact on the beneficiaries, while approximately a third of the cases indicated an impact level that might be considered as average.
With regard to the impact on the organization carrying out the social action we can see that the majority of them had an average impact, with a smaller percentage of actions having a weak impact. Actions with a strong impact for the companies carrying them out were identified in 36.3% of the companies we researched.
Joint analysis of the two items described immediately above suggests that the actions have a more significant impact for the organization than for the beneficiaries. With regard to the assumption about social responsibility, that companies do it because they have a greater consciousness of the needs of society, we would have expected that the percentages would be at least equal and it would be desirable for the benefits to be more important for the beneficiaries than for the organizations carrying out the projects. This suggests that the companies would need to redirect their projects, by taking these aspects into consideration.
With regard to the level of geographic impact we saw that this seems to be very concentrated, with 75.6% of the cases having an impact at the local level, 15.6% having a regional impact and only 8.7% having a national impact. These are values we expected to the extent that the majority of companies we researched were medium size, micro or small companies. In addition, it is relevant that the companies are focused on local development and that in this way they become development agents in this, their more restricted sphere of operation; it is desirable, but not obligatory that the scope of their operations becomes increasingly bigger as their actions improve.

Value generated
With regard to the value generated by the actions, and initially observing the social value generated, we were able to see that the majority of companies were involved with actions that were of low or medium social value, to the extent that only one in five companies (20%) were involved with actions that were highly valuable socially. It is worrying that a percentage of the actions put forward for the prize still generate low social value, since it would be desirable for the actions to have at least a medium level of social value. Those projects of low social value are the ones directed towards customer relationships, such as dances for the elderly, sporting competitions, etc, which despite being important, leave a lot to be desired when it comes to providing greater value for society, especially if we compare them to projects directed at street children, those with cancer and other similar actions.
With regard to economic value generated, we considered the level of projects of the companies analyzed really did represent an opportunity for generating economic value. From this point of view, the majority of actions have a medium or low impact.
It is important for companies that the social responsibility projects are converted into potential benefit for society and for the companies themselves, as a way of associating the benefit to society to the services provided by the organization carrying them out. What we found with regard to economic value generated is that it should also be at a higher level; the same goes for social value generated. The challenge for the companies, according with our analysis, is, therefore, to develop new projects with greater value, not only socially, but also economically.

existing collaboration
With regard to existing collaboration we initially tried to check if the projects were the initiative of the organization being analyzed or if they were being carried out in conjunction with some other organization, or were the initiative of another organization with which the company had become involved as a collaborator. The results showed that the vast majority of companies sought to develop actions on their own, with only 5% of the total indicating actions in which they were only involved as a collaborator.
With regard to partnerships, we tried to identify the existence of partnerships with non-governmental organizations, with government institutions or with some other type of institution. From what we could see the level of partnerships with NGOs and with government institutions is an auspicious, albeit restricted fact. This is because from what we could see there were partnerships with NGOs in 33.1% of the companies and with government institutions in another 23.8% of them. On the other hand, it is possible to see a large number of partnerships with organizations that are different from either of these.
The fact that various actions are being carried out in partnership with NGOs is good news, since the end purpose of these institutions is to develop actions in support of social development, which might be an indication of opportunities for NGOs and also an option of work for the companies. With regard to partnerships with the government it is possible to interpret the value found as reasonable (there were cases in which the companies had partnerships with schools, universities, city halls, etc.), suggesting that there are other possibilities for extending the benefit potential provided by these actions even further. The same can be said with regard to the number of partnerships with other partners.

leVels oF collaboration
With regard to the level of collaboration a total of 8 criteria were evaluated on a scale from 1 to 5, as shown in Figure 4 above. The measures and standard deviations are shown in Table 2. As can be seen, generally speaking the averages are fairly close to 2, being higher in the item, "Strategic value of the action" and lower in "Scope of the action activities". It is also noticeable that in all the items the standard deviation is close to 1, suggesting that in each case there are relatively few swings.
The first indication of these results is that companies concentrate more on the philanthropic and transactional categories. By adopting the measurement conventions proposed above and applying them to the tabulated data it was possible to identify that, in fact, 88.8% of the companies are categorized as philanthropic (in 36.9% of cases) and transactional (in 51.9%). Only 11.3% of the cases analyzed could be categorized as integrative organizations.

additional analysis oF data
To analyze the data we initially carried out an exploratory analysis, by looking at the correlation between the variables associated with the level of impact of the action, the value generated and the general evaluation of the results. The correlations proved to be high in the variables relative to the first two evaluation levels. We, therefore, moved on to factor analysis, using the Principal Component Analysis extraction method, with varimax rotation between the impact level variables of the action and the value generated. We clearly identified two factors, as shown in Table 3. From this table we can see that the variable "sphere of the initiative and of impacts" along with "social value generated" form the same factor, which we called "social benefit", and the two other variables form the second factor, which we called "economic benefit". The 8 social action general evaluation variables (levels of collaboration in Figure 4) were also submitted to factor analysis, but did not produce consistent factors.
We chose, therefore, to use the two factors generated from the analysis above for generating clusters. Analysis using the hierarchical method clearly indicated the existence of two clusters, so we moved on to an extraction, using the k-Means method. The description of the two clusters from the factors that generated them is shown in Graph 1 below. As is possible to observe: • Cluster 1 has higher scores in the two factors, social and economic benefit.
This cluster will therefore be called the group of high generated value. This group includes 98 companies, representing 62% of the total we analyzed; • The second cluster, on the other hand, represents lower scores than the previous one. This we will call the group of low generated value. This group comprises 61 companies, i.e. 38% of all the companies.
Proceeding to analysis by category of companies defined, as shown in Table  4 below, we can see that only the micro companies have a majority of their group that come under the cluster of low generated value. This can be explained from the association there is between economic size and the group of companies the company falls under. With the exception of shopping centers we can see that the larger the company the greater their level of participation in the group of greater generated value. On the other hand, in the analysis with regard to the level of collaboration, we can see first of all that the companies that carry out their projects without partners are equally divided between the groups with high and low value generation. When we look at Table 5 below it is clear that companies that want to be among those that generate greater value from their initiatives must develop partnerships. It becomes evident that the greater the level of collaboration, the greater the participation in the group of greater generated value. We analyzed the clusters by crossing them with the motivation for developing social actions. The results were as follows: • The significant majority (66.7%) of companies that develop their projects from a combination of social need and corporate capacity to help are in the group of those who generate greater value. The inverse occurs with companies that are not motivated by this. • The significant majority of companies that try to meet a relevant social need (72.5%) are in the group of those that generate greater value. Only 42% of the companies that are not so motivated are in this group.
• On the other hand, with regard to those that are motivated because of the interests of the executives, the majority, both of those that presented this particular motivation, as well as those that did not, were in the group of those that generate greater value. Even so, we can see a greater percentage of participation in this group of companies that are thus motivated than those that are not so motivated (65% and 51%, respectively). • The same conclusion can be drawn from analysis of the companies that are motivated by the public relations opportunity or value. • Finally, in the case of companies whose motivation is some form of government pressure, the two previous observations are repeated, albeit the participation of companies that have this motivation in the group of those companies that generate greater value (80%) is more pronounced.
In addition, we tried to check what would happen if we crossed the clusters generated with the geographic scope of the action. We saw that, regardless of the level (local, regional or national), the majority of companies are in the group of those that generate greater value. On the other hand it was possible to see that the greater the geographic scope, the greater the number of companies in this group, indicating that the vast majority of companies that develop regional (80%) and national (86%) actions are among those that generate greater value. As can be seen, the cluster of companies that generate greater value present higher scores in all 8 variables. This is not a surprising result, given that the companies that generate greater value should have higher scores in the evaluation variables.
On the other hand, two variables presented us with discordant behaviors. These are "level of breadth and scope of the activities" and "level of complexity of the management of the initiative". The explanation for this behavior must be because both the companies that generate greater value, as well as those that generate lesser value, may have broad, complex and difficult to manage activities, in other words, the level of complexity of the task may be also high in companies that generate low value.

coNclUSioNS aNd fiNal coNSidEratioNS
This study was structured using the concepts and proposals made by various authors and their application in the business practice of Brazilian retailers was checked. The results of this study suggest that use of this tool has made it possible to characterize three classes of company -philanthropic (36.9% of the cases analyzed), transactional (51.9% of the cases) and integrative (11.3%). From the analysis we conclude that when it came to social responsibility practices the majority of companies presented characteristics associated with the first two classes. In only 11.3% of the companies we analyzed did we see or identify characteristics associated with the integrative class, the most advanced of the classifications. These results suggest that, even though the retailers are already involved in important social actions and contribute to Brazilian society, there is an enormous potential for social contribution from these important agents.
We need to indicate some of the limitations of this work. First of all we used a database containing the social responsibility initiatives of retailers that report such initiatives for purposes of the prize award. In this initiative, which starts with the interested party putting itself forward and preparing a social report on what it does, we cannot ignore the possibility of the occurrence of the so-called social desirability bias (MOHR; WEBB, 2001).
On the other hand, it was the responsibility of the authors to analyze and judge the reports contained in the database by using the routine arising out of the bibliographic research. This option implies limitations, a fact that McDaniel Jr. and Gates (1999) drew attention to.
Another limitation we need to highlight is the application of concepts and dimensions conceived in other countries. In addition, the theme of social responsibility is a highly complex one and certainly involves a vision of the stakeholders with which the companies interrelate. The fact that this study is based on initiatives reported by retailers and for the purposes of a prize award is a considerable limitation, given the absence of the opinions of the stakeholders that are involved and interested in the reported actions. These stakeholders may have opinions and perceptions that are different as far as the cases contained in the database are concerned.
In this context of results and limitations we suggest future research into the theme of social responsibility. It would be desirable to consider the initiatives of retailers in contexts that do not involve a prize aspect. It is also be desirable in an effort to generalize the results that studies are carried out in other economic sectors.
In view of the fact that the initiative and database were set up in 2003 and the cases analyzed are of the transversal type, in other words, they happened at a particular moment in time and the social modifications intended occurred at different moments, it would be useful to carry out a study to determine the changes that happen over time.
Finally, the authors point out that despite the limitations imposed by the options and the methodological procedures of this study the results provide evidence that Brazilian retail companies are concerned with social issues and develop social responsibility actions in the course of their activities.