Introduction

In the past three decades, the concept of corporate social responsibility (CSR) has gained major attention, and today there is a common notion assuming that corporations have a certain, yet still somewhat undefined responsibility towards their stakeholders, especially the ones beyond consumers and shareholders. They are “our most powerful citizens and it is no longer acceptable that they be entitled to all the benefits of citizenship, but have none of the responsibilities” (Kercher 2007: 3). Most companies have realized this changing perception of the role of business in society, and started to establish or refine their CSR-strategies in order to satisfy this public expectation (cf. Utting 2005: 375).

In the course of this, corporations increasingly get involved in poverty reduction in the ‘Global South’. Here, they act as “development agents” (Volkert 2009) and implement CSR-strategies that appeal to poverty reduction. Knowledge about the effects of this engagement on the beneficiary groups, however, remains sparse: while various publications focus on the commitment of companies in poverty reduction, they often remain one-sided and many of them discuss the subject from a business-management perspective. Here, CSR is considered as a ‘business case’ and authors focus on the motivational background and the profitability of CSR-strategies. The title of a recent article gets right to the point: “Is doing good good for you?” (Lev et al. 2010; cf. Hamann 2007: 17; Carroll and Shabana 2010; Zadek 2004; Russell and Russell 2010; Porter and Kramer 2002; Schreck 2011). Another still emerging strand of research considers questions of impact and effects of CSR-strategies in the developing country context. It is, however, still in a nascent stage (e.g. Bolster and Brimble 2008; Nelson 2006; Blowfield 2007 or Werhane et al. 2010; cf. Prieto-Carrón et al. 2006: 986; Blowfield 2005: 523).

Jamali and Mirshak (2007) note that there has been some important success in the theoretical research about the effects of CSR in general, but the research on CSR and poverty in the South is still particularly lacking (Jamali and Mirshak 2007: 260; cf. Chapple and Moon 2005: 415). Yet, this takes the discussion to another level: “Assuming that CSR can be considered mostly a Western phenomenon, we can expect to have different scenarios when CSR is transferred to an extra-European context” (Wolff et al. 2009: 309; cf. Belal 2001; Hamann 2007; Jamali and Mirshak 2007: 260). Particularly research on CSR in Asian countries inherits this potential, since many of them are in a phase of transition and reveal most interesting perspectives for testing established “Western” theories (cf. Cheung et al. 2010: 401). Carroll (1999) notes: “For these [CSR-]concepts to develop further, empirical research is doubtless[ly] needed so that practice may be reconciled with theory” (Carroll 1999: 292). Despite the “prominent expectations that business can make significant contributions to sustainable development” (Hamann 2007: 25) and poverty reduction, yet we do not know “CSR’s consequences for the intended beneficiaries in whose name it is being conducted” (Blowfield 2007: 683) and whether CSR-strategies are indeed making a difference.

This article aims at contributing to this discussion of CSR’s impact on poverty reduction, focussing on the case of Cambodia. Two categories of CSR (built-in and bolt-on CSR) will be introduced in order to differentiate two types of CSR practices and to measure their respective impact on poverty reduction. Moreover, I will develop indicators for poverty reduction on the basis of Amartya Sen’s capability approach. Building on this methodological framework, the data collected in two case studies in Cambodia will be analysed, and the impact of two different CSR-strategies on poverty reduction measured. Both case studies are longitudinal analyses, therefore, an alteration of the variables can be observed. Since science “requires the additional step of attempting to infer beyond the immediate data to something broader that is not directly observed” (King et al. 1994: 7f.), in the final section, I will put forward some theoretical implications and identify further research potential.

CSR in development context as independent variable

The popularity, which the concept CSR has gained within the past decades, has not resulted in a conceptual clarification: “CSR knowledge could be best described as in a continuing state of emergence. While the [practical] field appears well established […] it is not characterised by the domination of a particular theoretical approach” (Lockett et al. 2006: 133; cf. Karnani 2011; Lee 2008: 54; Kercher 2007: 2; Maignan and Ferell 2000: 284, Clarkson 1995: 92ff.). Jonker states that “CSR is a ‘sensitising concept’: a term that draws attention to a complex range of issues and elements that are all related to the position and function of the business enterprise in contemporary society” (Jonker 2005: 20). In consequence, Okoye (2009) describes CSR as a contested concept with a problematic “lack of an agreed normative basis underpinning CSR practice […] and this has been linked to the absence of an agreed universal definition of CSR” (Okoye 2009: 614). Dahlsrud sees CSR as a social construction, which is why it “is not possible to develop an unbiased definition” (Dahlsrud 2008: 2). He identifies in a comparative analysis 37 different definitions of CSR, which sometimes overlap, but mainly stand in vast contrast to each other. Newell (2008) criticizes that “[w]hatever the reality might be of business-as-usual corporate business models, the rhetoric around commitments to sustainability, corporate citizenship and CSR has been deafening” (Newell 2008: 1064).

Understood literally, CSR refers to a general social (and/or environmental) responsibility of corporations.Footnote 1 However, especially regarding the scope of this rather unspecified form of responsibility, different CSR-definitions have always varied to a considerable extent. While Friedman (1970) exclusively reduced CSR to the mere economic responsibilities of corporations (“the business of business is business”), Davis (1973), for example, stated that “CSR requires consideration of issues beyond the narrow economic, technical, and legal requirements of the firm” (Davis 1973: 312). Even today, in many CSR-related articles the actual scope of this responsibility is only marginally addressed. Instead, there is often an emphasis on the voluntariness of this vague form of responsibility. Until October 2011 for example, the European Commission stated in its Green Paper ‘Promoting a European framework for corporate social responsibility’ that CSR is “a concept whereby companies decide voluntarily to contribute to a better society and a cleaner environment” (EC 2001; emphasis added). Piacentini et al. (2000) similarly argued that “CSR is the voluntary assumption by companies of responsibilities beyond purely economic and legal responsibilities” (Piacentini et al. 2000: 459) and van Marrewijk (2003) stated that CSR is “voluntary by definition” (van Marrewijk 2003: 102). The weaknesses of this understanding of CSR, however, are quite obvious: on the one hand, it is questionable, whether CSR measures can per se be voluntary, because the boundaries between voluntary and obligatory regulations vary from country to country and are often rather blurry (Blowfield and Frynas 2005: 502f.; Moon and Vogel 2008: 265). Just recently, the European Commission seems to have started to accept that the social responsibility of corporations needs to go beyond voluntary commitments of corporations. In the new CSR-communication of the European Commission, CSR is defined as “the responsibility of enterprises for their impacts on society” (EC 2011), which provides a much broader definition of CSR than before, refraining from the explicit focus on the voluntariness of CSR. The other weakness of the current definitions of CSR is the unanswered question of what this responsibility actually comprises, which is why many of the definitions of CSR can be criticized as being rather arbitrary.

“The frustrated search for a precise definition with demarcated categories has left ambiguous how different activities […] all might be misunderstood in terms of pertaining to CSR. This lack of a tight conceptual centre has caused many to critizise CSR, and even to ask if this is really best understood as one concept or instead as rather an umbrella term […]” (Sabadoz 2011: 77f.). In consequence, this article will define CSR as a responsibility that involves all affected stakeholders, beyond a narrow fixation on consumers, shareholders and primary stakeholders, especially integrating stakeholders of the periphery of the supply chains or in the direct (physical) surrounding of the company’s activities. CSR can be a “serious attempt to solve social problems caused (…) by the corporation” (Fitch 1976: 38; emphasis added) or “about creating acceptable wealth and distributing it to a growing number of stakeholders in a correct and justifiable manner” (Jonker 2005: 20). For heuristic purposes, I understand CSR here as a concept where any kind of company either conducts its core activities in a way that creates considerable positive (or at least not negative) effects for their primary and secondary stakeholders on a quasi-voluntary basis (cf. Fitch 1976: 38), or as measures by which companies voluntarily go beyond their core business and take some sort of action in order to create positive effects for stakeholders apart from consumers and shareholders, often focusing on stakeholders that are not part of the direct value chain, e.g. the communities in which the company operates (cf. Jonker 2005: 20).

This distinction is also referred to by Barth and Wolff (2009). In their approach, the CSR-activity is either built-in to the core business operations or it is bolt-on to the regular business processes, and can hence be labelled as philanthropy. I will use this distinction in order to distinguish different degrees to which CSR is integrated into the core management of firms. When more integrated, “CSR is about building social (…) concerns into a company’s operation” (Barth and Wolff 2009: 14) and directly into its core business activities. They are therefore built-in to the core business and oblige corporations to do what is “right, just and fair” (Matten et al. 2003: 110) even when the business is not compelled to do so by a legal framework. When CSR is less integrated, the company’s CSR-activities lie outside a firm’s immediate core business operations, and are thus bolt-on (Barth and Wolff 2009: 14). Here, companies engage in individual initiatives beyond their core business activities (Hiß 2005: 38 and Barth and Wolff 2009: 14). Although bolt-on CSR is often strategically related to the core business, it usually does not interfere with the company’s typical business operations (Barth and Wolff 2009: 14).

Poverty as dependent variable

The first of the Millennium Development Goals is dedicated to the eradication of poverty: the number of people living in extreme poverty shall be halved by 2015 (See MDG website). But what does the term poverty mean exactly? In the past, poverty was exclusively defined as income poverty. The growth rate of the gross domestic product (GDP) was over decades the only indicator for poverty, and the argument of a trickle-down effect of wealth to the poor was brought up “to defend such neglect” (Srinivasan 1994: 238). The understanding of poverty in this article is based on the Human Development and Capability Approach by Amartya Sen, which is also being used in the Millennium Declaration and the Human Development Report (Sen 2000b: 17). Throughout the last 30 years, Sen has developed and refined this framework which approaches development with regard to human capability and freedom (see amongst others: Sen 1985, 1992, 1999, 2000a, b). His work is among the most important in the field and has been taken up and further developed by other scholars. Following Sen, poverty is more than mere shortage of income and is defined with respect to a lack of human development (Volkert 2009: 393) which includes “multiple forms of deprivation and social exclusion (…), often based on gender, (…) ethnicity and class inequalities” (Newell and Frynas 2007: 673): “Wealth is evidently not the good we are seeking; for it is merely useful and for the sake of something else” (Sen 1990: 44). Therefore, eradicating poverty means to create an environment which enables people to develop their full potential and conduct a life that is consistent with their needs and requirements: “Development is (…) about expanding the choices people have to lead lives that they value” (Sen 1997b: 1959) and does not put people’s self-respect into question (Sen 2000a, b: 37). When discussing poverty, we should not (only) consider what people have, but if they can achieve what they want to achieve and recognize the difference between what people are able to be/do and the means they possess to follow this path (cf. Robeyns 2005: 94). In this view, individual wealth should not only be characterised by income or primary goods, but in terms of the ability to choose freely between different ways of living. Poverty reduction is thus much more than pursuing economic growth.

Based on Sen’s work, the term poverty in this article will be defined as a lack of substantial liberties to live a self-determined life, thus, a lack of capabilities (Sen 2002: 110) which comprise abilities, conditions and realization chances and require a certain amount of effort for alteration. Every individual has his or her own life plan which he or she wants to perform. Each of these life plans is in need of a certain combination of capabilities, the so-called capability set (Sen 1992: 39ff.). This set describes a bundle of capabilities that a person is in need of in order to fulfil his or her life plan (cf. Volkert 2009: 394; Robeyns 2005: 98 and Srinivasan 1994: 239). According to Sen, capabilities reflect a person’s opportunities and the freedom of choice between different possible life plans (cf. Sen 1985, 1992, 1999). This capability set “stands for the actual freedom of choice a person has over the alternative lives that he or she can lead” (Sen 1992: 114). Reducing poverty must therefore mean expanding all basic capabilities so that the chance increases that the poor can obtain their very individual capability set. It is, however, important to notice that “it is the people directly involved who must have the opportunity to participate in deciding what should be chosen, not local elites […] or cultural experts” (Sen 1999: 31f.). Bearing this in mind, ten different factors will be extracted from the relevant literature in order to create an instrumental outline for measuring CSR’s impact on poverty. However, this list does not claim to be complete and the capabilities in it can be interpreted rather widely:

It is unquestionable, that income is still one of the most important means to increase capabilities regarding poverty (e.g. Volkert 2009:10). Despite the proclamation of a wider sense of the term of poverty, it will also be the first factor in this list. “Income is, of course, a crucially important means, but its importance lies in the fact that it helps the person to do things that she values doing and to achieve states of being that she has reasons to desire” (Sen 1997a, b: 385). The level of education is a capability that applies to both the earning of income and converting this income into life improvement. It is one of the largest determinants for future income: with each finished school year, the average income increases by 10% (Burnett 2008: 270). A lack of education is therefore both a handicap for income-earning and income-using (Sen 1992: 113, 1997b: 1959 or 1995: 14), because it gives the people “knowledge and tools to break the cycle of poverty” (Burnett 2008: 269). The lack of the third capability, the empowerment of women and youth is one of the major problems that obstruct poverty reduction: “Women [as well as the old or the young] may have special disadvantages in converting income into particular functioning” (Sen 1992: 113; 122ff.) which would allow them to live a self-determined life (cf. Sen 1992: 113, 1997b: 1960). The next capability is individual health. This has been in use besides the factor of income for long, even at times when the growth of the GDP per capita still was the sole indicator to measure development (Srinivasan 1994: 238). A life plan can only be fulfilled consequently, if the current health status permits it (Sen 1995:14, 1997a, b: 386). Accommodation or enjoying a “social” environment (Sen 1997a, b: 386) is the fifth capability which is crucial for individual poverty. The well-being of the individual depends on whether the place of domicile is strategically well-located (Sen 1992: 113) and provides adequate assets in order to fulfil the individual life plan. The sixth capability is the access to the labour market and controlled labour conditions (cf. Volkert 2009: 387) which refers to the importance of participating in working and production processes. Unemployment or poor labour conditions do not only lead to low incomes, but also to non-monetary consequences, like health problems or social exclusion. In close connection with this is the access to capital, land and insurance. An improved access to capital and credit enables poor people to invest and achieve sustainable improvements, e.g. to found a small business (Volkert 2009: 398). An improved access to insurances enables the individual to focus on these sustainable improvements. The degree to which illnesses affect the well-being of individuals depends on the access to a functioning health care infrastructure (Volkert 2009: 397). The possibility to be medically treated simplifies health improvements, and enhances the individual’s possibility to focus on the actual fulfilment of the life plan. The access to social networks and the possibility to participate in political and societal activities create a further foundation to improve an individual’s life situation not in monetary regards, but with regard to identity and appreciation (Volkert 2009: 399). Political freedom and the right to participate in political processes opens possibilities to control and criticize the state (Volkert 2009: 402), which can enhance satisfaction and social involvement. The tenth capability of social security and the access to a regulated legal system is essential to face (monetary) emergency and/or a criminal environment (Volkert 2009: 400). Only with a certain amount of security, it is possible for the individual to invest in the fulfilment of the life plan in a sustainable manner (cf. Haq 2000: 77f). Sen points out that for some people, the commodities of social security and features related to community living, such as social cohesion, the prevalence of crimes and mutual help exceed the need for financial commodities and resources (Sen 1995: 9).

Each of the capabilities demonstrates a possible starting-point for poverty reduction. Therefore, a positive impact on one single capability can ignite a process of poverty reduction. Regarding the measurement of poverty reduction, the alteration of one single capability is a required minimal impact. Of course, the more capabilities are positively affected by CSR-strategies, the bigger the impact of this process.

Hypotheses about the impact of CSR on poverty

  1. H1:

    CSR has direct impact on poverty

    “Private enterprises offer the potential to (…) improve the quality of life for millions of people around the globe” (Nelson 2006: 41). The first hypothesis assumes that CSR has a direct positive impact on poverty (cf. Volkert 2009: 406; Blowfield 2005: 516). The British Department for International Development also confirms this assumption: “By following socially responsible practices (…), the private sector will be more inclusive, equitable and poverty reducing” (cited in: Jenkins 2005: 525 and 530; emphasis added). With this first hypothesis, I agree with Chesters and Lawrence (2008) and Idemudia (2008) who stress the connection between CSR and poverty and even indicate that although this connection has not been investigated enough, it is indisputable that it exists. “Almost all companies, regardless of industry sector and other variables have the potential to make a sustainable contribution to (…) poverty alleviation” (Nelson 2006: 47). Besides this, many enterprises are beginning to put up their social activities explicitly with the main goal of fighting poverty (Boyle and Boguslaw 2007: 102) which reveals the relevance of the subject to companies. The relevant literature affirms a positive correlation of CSR and poverty reduction. “[T]here is a wide variation in the specific contributions that different companies can make (…) to supporting key development goals such as poverty alleviation” (Nelson 2006: 46).

  2. H2:

    Built-in-CSR is more effective than bolt-on CSR

    “The greatest and most sustainable contribution that any company can make to development is through carrying out its core business activities in a (…) productive and responsible manner” (Nelson 2006: 47). The second hypothesis claims—similar to the quote—that built-in CSR has a larger positive impact on poverty than bolt-on CSR. Schuster (2005) confirms this by saying that the potential to attain a significant impact on the achievement of the Millennium Development Goals is much higher when the core business activities are conducted in a socially responsible way, hence, when CSR is coherently built into the usual business operation (Schuster 2005: 63). “99.9% of our impact (…) is through our business activities. Even if we stopped the philanthropic programs, we’d still be achieving 99.9%” (Newell and Frynas 2007: 671) and “[s]ustainability is impaired, when companies do not regard the tackling of a CSR issue as strategically important” (Wolff et al. 2009: 295f.; cf. Hamann 2003: 239 and 243). It therefore seems crucially important for the effectiveness of CSR if the issue is strategically important to the company’s own business interests. Another theoretical argument for this hypothesis is the link of corporate responsibility with the direct stakeholders (the ones that are affected by built-in CSR). This responsibility is higher than the responsibility for stakeholders outside the core business (the ones that are affected by bolt-on CSR). There are thus more reasons for a company to assess the impact of its built-in CSR strategy more carefully in order to reach the goal of stakeholder satisfaction. On the other hand, by implementing bolt-on CSR only, the company does not consider the effects on stakeholders as much. Therefore, the company will try to enhance the impact of its built-in CSR-activities, since it is exactly what satisfies the direct stakeholders. In bolt-on CSR, the actual impact on the stakeholders (which are not part of the core business) does not seem to be as crucial and in many cases even unassessable. Here, it is just important that any CSR is being conducted, so the company can improve its reputation. Wolff and others also follow the same argument. They claim that when impact is more difficult or less important to measure (e.g. bolt-on CSR) it is more demanding for companies to assess this impact and review its progress. “Impact is more difficult to achieve when the issue lies outside the company’s immediate sphere of influence” (Wolff et al. 2009: 297).

Methodological approach—a framework for assessing CSR’s impact

In order to test the hypotheses, a research design will be established with which I can identify impact of CSR on poverty. In order to verify the research results, “a set of rules of inference” (King et al. 1994: 9) will be elaborated in this chapter. This set includes indicators with which I can identify impact. The empirical data has been conducted in two case studies in Cambodia. The first case study includes the involvement of over 1,000 small corporate entities in poverty reduction of street children in Phnom Penh. Their activities have taken place external to the corporations’ core business and are therefore bolt-on. The second case study comprises the involvement of large transnational textile companies in improving working conditions in Cambodian garment factories. In this case study, CSR takes place within the core business of the companies and is therefore built-in. The selection of very small businesses in the first and very large companies in the second case study is justified by the “U-shaped” (Udayasankar 2008: 173) effect of firm size in CSR participation. According to Udayasankar, it is mostly the small or large firms that apply CSR-strategies of which one of each is considered in this study.

In order to analyse potential changes that have been brought along by CSR, the process of the CSR-activities will be split up into three sections. For each case, the different CSR-strategies (the CSR-input) will be described. After this, the outcome will be analysed before the impact can be assessed. The CSR-input, in this article, refers to the change of behaviour of corporations as a result of a changed CSR strategy. Outcome refers to the concrete changes that are taking place as a result of this change of behaviour. If, for instance, the CSR strategy of a company includes the improvement of health and safety issues, one of the outcomes of this decision would be the installation of fire alarms or the training of workers. The outcome can immediately be experienced by the target group whereas the CSR-input is a behavioural change of the company.

Impact in this study technically means “implications for society” (Skajærseth and Wettstad 2009: 33). It comprises the direct and indirect positive effects that have been brought along by the businesses’ CSR-strategies and the resulting outcomes. It should not be mixed with outcome, since it refers to actual life changes of the target group. In order to analyse this impact, I will consult the capabilities for poverty reduction again. They build the foundation on which changes can be observed. Sen ensures that in evaluating impact on his approach of poverty, the change of each capability will be different. They can, however, be measured equally. Therefore, after identifying changes of the capabilities, they will be sorted into an impact matrix, although these changes might differ according to the capability. The focus lies on “relative improvement” (Barth and Wolff 2009: 29).

For the analysis, indicators for the capabilities have to be found with which I can confirm that a change has taken place. The following table will list a range of indicators that have partly been extracted from the literature. For each capability, two indicators will measure potential changes: one indicator for the direct, rather immediate impact of the outcome and one indicator for the indirect impact which is evidently slower. The reason for choosing two indicators is the following: with only one indicator, the difference between direct and indirect impact could not be distinguished. With more than two indicators, it is more difficult to guarantee that all indicators exactly measure the same, which is essential for the impact analysis (Table 1).

Table 1 List of indicators (direct and indirect) that relate to the capabilities
Full size table

With the help of the following matrix, it will be possible to identify impact in the two cases. It is important to state, however, that the analysis remains qualitative and the matrix exists to get an overview of the impact. In order to categorize the impact of each case, three routing questions will be answered for each of the ten capabilities and the answers rated on a scale from 0 to 4. These results are then added up for each capability which will give an impression of the quantitative scope of the impact on each capability.

Routing questions:

  1. 1.

    How many indicators can be confirmed and are they direct or indirect?

    The impact increases with the number of indicators that have been confirmed. If the direct indicator of a capability can be confirmed, the impact on this capability can be identified comparatively clearly. If the indirect indicator can be confirmed it is more likely that there will be a further improvement in the future, which is however more difficult to measure at present.

  2. 2.

    What is the duration of the impact?

    Impact can either be short-dated, long-term or both. Short-term impact is important, because it demonstrates quick changes which influence the individual’s life immediately. Long-term impact is equally important, since it influences the individual’s life in a longer perspective. The largest impact is certainly a mixture of both, because it changes lives immediately and is still sustainable in the long run.

  3. 3.

    How can the impact be attributed to the influence of the corporate involvement?

    In order to refer impact back to the involvement of business, it is important to trace back the influence. It is generally difficult to attribute changes at the end of the results chain to its beginning because this concerns comprehensive and complex changes in the environment of the project, in which several factors are involved. It is methodologically challenging to clearly attribute the higher level changes to the contribution of an individual project (Reuber and Haas 2009: 5), and it is questionable that these changes were really induced by the corporate involvement. Therefore, the bigger this “attribution gap” is, regarding each capability, the lower the rating of the third question will be. In order to analyse the quantitative impact of the CSR-strategies, the answers to each of the three routing questions will be added up, so that an overall impact on each of the ten capabilities (measured in units from 0 to 12) can be demonstrated (Table 2).

    Table 2 Routing questions and categorization of the answers into an impact scale
    Full size table

Empirical effects of bolt-on CSR on poverty: ChildSafe

The first case study focuses on the cooperation between the program ChildSafe of the international NGO Friends-International and more than 1,000 small and smallest businesses. This first case involves three different types of stakeholders: first, there are the street children as target group of the program. Secondly there is the ChildSafe program and Mith SamLanh as main referral partner. Then, there are the so-called ChildSafe Members which consist of private businesses that are certified to be ChildSafe.

According to Terre des Hommes, there are approximately 1,200 street children in Phnom Penh. According to UNICEF and Friends-International this number even ranges between 10,000 and 20,000 (Köstler 2006: 6). The number depends on the definition of street children and the time of the year. The majority of the street children are male and most of them have very insufficient formal educational levels.Footnote 2 The reasons for the high number of street children differ widely, ranging from agricultural failures and health problems to domestic violence and poor access to resources. One of the major reasons, however, is the massively growing tourism sector in Cambodia. The sector has experienced a growth rate of about 20% in the past decade which will remain constant over the next years. This growth is accompanied by a fluctuation of people to urban regions because of their expectation to find employment within this sector. This mass of the so-called inland migrants consists mostly of youths and children who are exposed to an extreme vulnerability due to their young age and neglect. The problems that street children have to face cover a wide range and can only rudimentarily be listed here: violence, drug abuse, health problems, mental and emotional disorders, (sexual) exploitation and trafficking.

ChildSafe is one of the programmes of the international NGO Friends-International and was founded in 2005. The main goal of the program is to create a safe environment for street children. It advocates societal responsibility by involving local and international businesses in the protection of children. The co-program of ChildSafe is Mith SamLanh, an affiliated centre in which different trainings and psychological assistance are offered and outreach programmes are being conducted. The largest ChildSafe project is the cooperation with the ChildSafe members, the so-called ChildSafe Network. These ChildSafe members consist of over 1,000 small and smallest private businesses which predominantly work in the tourism sector. These businesses are certified to be ChildSafe, following a clear process lead by the ChildSafe centre. The ChildSafe members are trained to recognize children who are at risk and become actively involved in their safety. Regular contact, supervision, monitoring and follow-up with the ChildSafe businesses ensures that the ChildSafe policy is respected and implemented. Almost all ChildSafe members benefit financially from the reputation of the network. According to a survey conducted by the ChildSafe team, taxi-drivers had their income increased by an average of 30% after becoming ChildSafe certified. So far, 1,330 Cambodian businesses have adopted the ChildSafe policy, of which about 1,100 are located in Phnom Penh. On average, this number has doubled every year since its establishment. After the first contact with the child on the streets (by one of the businesses), most children are referred to Mith SamLanh where they can take part in training programmes or other educational services. The businesses together with the ChildSafe centre are therefore the main entrance channel to Mith SamLanh for Phnom Penh street children. Every registered child gets a chance for life improvement. The cooperation with the private businesses is essential for recognition, identification, emergency aid and the referral of the children.

The main reason for choosing the ChildSafe program as a first case study is the localisation of the corporate involvement outside of the core business: the protection of street children and the creation of a secure surrounding for them are not involved in the business planning of any of the businesses. This offers the opportunity to conduct research within the bolt-on CSR. In this case the missing link between core business and the recipients of the CSR-activity is even stronger than in other cases: the target group of the program—the street children—is by no means a potential customer group or in other ways involved in any part of the businesses. This emphasizes the affiliation of the case with the bolt-on CSR in which—by definition—the stakeholders are located outside the core business operation. In the regarded case, the businesses apply their CSR-activities together with an NGO, which gives the CSR-activities an institutionalized framework. The argument of Jamali and Mirshak, who claim that the partnership of business and NGOs with regards to CSR are “pure outsourcing exercises” (Jamali and Mirshak 2007: 259), can be disproved here: without this institutionalized framework, a similar network to the ChildSafe Network would not have been established as easily. Thus, the cooperation of the private sector with an NGO is essential for this kind of CSR-activity. Also, as Azmat and Samaratunge (2009) argue, the little impact assessment about CSR-initiatives has been conducted only with large multinational companies. “Given the important role of SIEs [small scale individual entrepreneurs] in developing countries in (…) alleviation of poverty, their social responsibility deserves considerations, but remains severely under researched” (Azmat and Samaratunge 2009: 438; cf. Moore and Spence 2006: 223; Brian Lund in the CSR Conference Report 2008). This is another reason for the case selection.

Analysis of research results

In order to analyse the impact of the program the main CSR-activities of the involvement and the outcome have to be summarized: businesses that are members of ChildSafe discover children in need or at risk and provide emergency aid, the businesses can bring street children to the ChildSafe centre which is open 24 h every day, the businesses and the centre refer the children to Mith SamLanh and other referral partners and the businesses that are ‘childsafe’ maintain presence around the city of Phnom Penh.

The drafted activities lead to an outcome with different aspects: the first main outcome is that children have a place of refuge and are therefore provided with nutrition, psychological support and stress relieve. Furthermore they are able to get (health) care and immediate medical or psychological support. They can meet fellow street children in a similar situation and they are able to gain more knowledge in a variety of different aspects. For many of them it is the only place where they can learn how to read. Stability and regularity is created for the street children. The Phnom Penh population is more aware of the existence of ChildSafe and potential criminals are aware that they might be discovered by civil society, e.g. the ChildSafe members. ChildSafe business owners start to act “ChildSafe” outside of their work place as well. There has been evidence, that many CS members are helping children outside of their usual workplace, for instance in their neighbourhood. About 700 times a year, the ChildSafe hotline is being called to report the abuse of children. About one third of the callers are street children themselves, the others are mostly ChildSafe businesses or individuals that have recognized children in need (Table 3).

Table 3 Overview of CSR-input and first outcome of the involvement of CS
Full size table

For the actual impact assessment, I will analyse the effects on every indicator of the ten capabilities for poverty reduction. The changes regarding each capability will first be described and then sorted into the impact matrix:

The capability of income has not been tremendously affected by the involvement of ChildSafe, since the program does not generate income. However, the program does provide the opportunity to be trained and qualified for a job that is going to pay much more income than without the qualification. After absolving a training successfully, most of the former street children find a profession which provides them with a secure income. There is evidence that about 60% of the students that have started training at Mith SamLanh finish it successfully and a smaller percentage finds a job in the relevant field. Some of the children that are addicted to drugs go through one of the offered detoxification programmes to which they are being referred to by ChildSafe. Therefore many of the former drug users can fight their addiction and spend less money on drugs. Since the drug use of many of the children was one of their main expenses, they can use their income in other ways, which gives them relatively more money. Therefore the first indicator has not been confirmed, whereas the second indicator has. Furthermore, this impact is certainly only noticeable in long-term and it is clear that the perspective for increased income can be attributed back to ChildSafe, because only by the involvement of the businesses, this referral to the centre and the training at Mith SamLanh could take place.

In the centre of Mith SamLanh, 700 children either go to school every day, or take part in one of the eleven training and education programmes. Therefore, there is apparent impact on the level of education of the street children which are affected by the program. It is evident, that they have gained more knowledge than before and the average illiteracy rate has decreased considerably. There is also an increased perspective in raising this knowledge in the near future due to the long-term endurance of the training programmes and the possibility for street children to take part in the programmes as long as they want. One of the main reasons in Cambodia not to attend school is the school fees which are raised in most schools, despite being illegal. Another main reason for not attending school is that children have to look after their smaller siblings while their parents earn money (Köstler 2006: 143). Mith SamLanh, the affiliated centre to ChildSafe, has eliminated both reasons by providing the trainings free of charge and by providing child care centres. The impact on the level of education is strong on a long-term perspective, because of the long-running educational programmes. However, there is also a slight short-term effect, since the children are immediately able to take part in the training programmes and have the chance to increase their knowledge instantly. They are also able to get a training certificate after the first three training months (Köstler 2006: 31). It is clear that the enhanced level of education can be attributed back to ChildSafe, since it refers most of the street children to Mith SamLanh. However, there are also different ways to enter the same training programmes; therefore the impact on the level of education can not significantly be attributed back to the involvement of the ChildSafe program.

Many of the educational training sessions focus explicitly on female topics and the empowerment of women, the next regarded capability. One of the focuses lies on the involvement of women in the labour market, and possibilities for girls to earn money apart from sex business. Girls and young women are also supported by being offered the same opportunities as male children, which is very uncommon in Cambodia, and by being offered a place to go to in case of emergency (sexual harassment, early pregnancy, etc.). There is sufficient empirical evidence that there is enhanced safety for young girls to avoid rape, sexual harassment and trafficking on the streets. The name of the network implies the focus on young people and youths. ChildSafe focuses specifically on the empowerment of youths and children, therefore, the importance of the younger generation is underlined in a society where children are still not of much worth and only exist on the edge of society. However, there has not been a noticeable effect on the discrimination of youths (and women) outside of the centre and the NGO. There is a possibility that the children are more aware of their rights and have increased self-confidence due to the business involvement and the training programmes. Yet, this is not at all confirmed and can therefore not be included in the research results. Hence, the indirect indicator can, whereas the direct indicator cannot be confirmed. The impact on girls and youths in general is strong from a long-term perspective, because the level of education and awareness increases with time. The changes regarding this capability can be attributed to ChildSafe, since it is the major program in Phnom Penh to tackle this issue and focus specifically on youth. The increase of security standards for girls on the streets is a direct effect of ChildSafe, since it is the only program of its kind.

The impact of ChildSafe on the individual health level of the target group is visible, although only on short terms. Emergency health care and aid is provided by ChildSafe, however, long-term health care and the follow-up of the patients is not possible. There is no evidence that there have been fewer diseases after the involvement of ChildSafe. Hence the direct indicator cannot be confirmed. However there are several health education programmes offered by ChildSafe. Sexual education is also an important matter in many of the programmes of Mith SamLanh and ChildSafe. Street children are empowered to take part in these programmes in order to prevent sexually transmitted diseases. The impact is therefore rather indirect, but it can be attributed to the involvement of the ChildSafe businesses and ChildSafe, since it definitely provides first-aid and emergency shelter. Furthermore, the businesses provide a certain health security on the streets, since they are trained to recognize and help children with health emergency: when a ChildSafe member recognizes a child in danger, the centre is contacted which—together with the business—provides the child with the necessary measures, e.g. medical check-ups or the referral to a hospital.

There is some impact on the capability of accommodation since there are certain possibilities for street children to be accommodated in boarding houses. There is also the possibility to stay at emergency shelters for a few days. However, the overall domicile situation of the children has not changed, since most street children that were in touch with ChildSafe or Mith SamLanh still live on the streets or in simple accommodations around the city. There might therefore be a certain impact of ChildSafe on these children which seem to be more concerned about their accommodation. However, the direct housing improvement is only visible in the short term, and longer changes in the attitude towards accommodation cannot clearly be attributed back to ChildSafe. Neither the direct nor the indirect indicator can be confirmed regarding this capability. Nevertheless, there still is a slight impact on this capability since the children have a chance to be accommodated a little better for a short time period.

There is no direct impact of ChildSafe on labour conditions. Still, the access to the labour market is considerably improved by ChildSafe due to the possibility of taking part in the vocational training courses at Mith SamLanh. Former street children that have completed a training course are much more likely to work with good working conditions and better payment than without the trainings. However, there is no evident impact of ChildSafe on the improvement of labour conditions, and there are no specific training programmes that are being offered for the awareness about labour conditions. However, the existent impact can be attributed back to ChildSafe, since it is clear that it is the training courses of Mith SamLanh that enhance the access to the labour market. There is evidence, that without these trainings, the children would not have been employed. The follow-up of the cases shows that most of the former street children are able to live on their own and have a regular secure job. However, since the impact is very indirect, it is difficult to say whether the children would have had other chances to be employed in the same occupation.

There is no evident impact on the access to capital, land and insurances. The perspective of increasing income might influence the ability to get a loan, purchase land or procure insurance. However, there is no empirical basis for this assumption. Neither the direct, nor the indirect indicator can be confirmed.

The impact on the access to health care system is ambivalent. On the one hand, it is confirmed, that members of the target groups go to doctors and hospitals and are medically checked up more frequently. The ChildSafe centre has referral partnerships with several hospitals around Phnom Penh and through these partnerships, the discovered street children in need can get appropriate health care and hence access to the regular health care system. On the other hand, this impact is only existent in short-term perspective, since the long-term access to the system cannot be ensured by ChildSafe. Therefore the impact is very small, since it does not have any effect on the systematic structure of the Cambodian health care system. It, however, provides a certain systematic structure for the street children to access the regular structures of health care. Therefore the direct indicator can, whereas the indirect factor cannot objectively be confirmed.

The impact on the access to social networks is most likely the strongest instrument of ChildSafe. Together with Mith SamLanh, it provides a network-like social basis for street children where every day about 700 street children can attend educational trainings and about 1,800 children are worked with during the outreach programmes on the streets. In summary it can be said, that the impact of ChildSafe on the creation of social networks is comparably high. These newly created networks build a safe surrounding for the street children and provide a substitution for the networks of street gangs and others. There is, however, almost no possibility for the children to be heard by media and others. The public awareness does not rise significantly because of these networks. The indirect indicator, thus, cannot be confirmed. It is nonetheless evident, that a major strength of ChildSafe and Mith SamLanh is the building of these networks. Creating new types of networks also consolidates the success of the program and shows the street children, that there is a possibility to get support. In short- as well as long-term perspective, children can take part in the networks.

Regarding security and the access to legal system, the largest contribution of ChildSafe is the prevention of dangers for the street children. The well-known logo and the reputation of the ChildSafe businesses work preventive for any prospect violent or sexually abusive people. The ChildSafe centre provides help for children that have been victims of violence from street gangs, the so- called Bong Thom. Children in danger are noticed by the ChildSafe business and are reported back to the ChildSafe centre. There is no measurable impact of ChildSafe on the access to the legal system in Cambodia. Therefore, the impact on this capability is definitely measurable, but only for extreme situations of endangered security. Most of the street children within the city of Phnom Penh have heard of the centre and therefore have a place to go to when they feel endangered. It is evident, that around the area of the ChildSafe centre and Mith SamLanh, there is less danger for street living children. There have been several cases in which a dangerous situation for a street child could be relieved by one of the businesses. The direct indicator can therefore be confirmed. This small but measurable impact can clearly be attributed to ChildSafe, since it is the only program that works together with regular “non-street-worker” people from the Cambodian society (= the businesses). There is no measurable impact of ChildSafe on the access to the legal system in Cambodia. Therefore, the impact on this capability is definitely measurable, but only for extreme situations of endangered security (Tables 4 and 5).

Table 4 Classification of the capabilities into the analysis scale
Full size table
Table 5 Ranking of impact of ChildSafe regarding different capabilities
Full size table

Empirical effects of built-in CSR on poverty: BetterFactories Cambodia

The second case study takes the cooperation between BetterFactories CambodiaFootnote 3 (BFC), the international textile buyers and the Cambodian garment factories into closer consideration. It will focuses on the social benefits of the BFC-program, especially whether the change in strategy of the international textile buyers has had a positive impact on the Phnom Penh garment workers regarding the ten capabilities for the reduction of poverty. This second case study involves four different categories of stakeholders: again, there are the recipients of the program, which are the predominantly female garment workers in the textile factories. Second, there is the program BetterFactories Cambodia of the International Labor Organization. Then there are the international buyers, and the Cambodian garment factories.

In Cambodia, there are about 400,000 garment workers, working in more than 500 garment factories. Most of them, about 320,000, work in and around the capital of Phnom Penh. This number depends on the orders of the international textile buyers and varies also according to the seasons. The ordinary garment worker is female (over 90%) and between the age of 18 and 24. Most of the workers are originally from one of the Cambodian provinces and came to Phnom Penh in order to find work to support their families.

The involvement of the international buyers in the textile sector in Cambodia has changed within the past years. The attention of the big international clothing companies regarding the producing conditions of the imported clothing has increased during the last years, because of public awareness, media attention, civil society pressure and the coherent enforcement to set up CSR-strategies.Footnote 4 Adidas now states in its sustainability report, that it is one of the company’s biggest concern to integrate reactions and opinions of stakeholders like the textile workers into the improvement of the supply chain. An improvement of compliance regarding social standards, extended communication with stakeholders and a progress in monitoring mechanisms and standardisation of labour conditions are some of the fields that have been worked on by the international textile companies within the last years (Adidas Sustainability Report 2008: 3ff.).

The growth of the Cambodian garment industry found its beginnings around 1996, 3 years after the final end of the civil war. After the reestablishment of the Kingdom of Cambodia in 1993, the first (foreign) investors to venture into the country were the garment manufacturers. On January 1999, the Governments of Cambodia and the United States entered into a Trade Agreement on Textile and Apparel. This agreement set an export quota for garments from Cambodia to the United States, while seeking to improve the conditions for workers in the garment sector. The Agreement offered a possible 18% annual increase in Cambodia’s export volume if a program was established to improve the conditions for garment workers. The result was the program Better Factories Cambodia which is made possible by the International Labour Organization.Footnote 5 It is closely working together with the garment factories in Cambodia and the international buyers through which—amongst other sources—it is funded. The basic objective of this program, according to its website, is to “benefit[s] workers, employers and their organizations. It benefits consumers in Western countries and helps reduce poverty in one of the poorest nations of the world” (BFC Homepage). With Cambodia’s entry to the World Trade Organization in the year 2005, the Cambodian garment sector is now acting on a free market and has to face the international competition after years of economic protection by the USA and European countries. The garment industry has been growing steadily until mid-2008 and has maintained its pre-eminent position in the industrial landscape of Cambodia. Today, Cambodia is one of the ten largest suppliers to export clothing to the US and Europe.

In October 2008, 285 factories in and around Phnom Penh were part of BetterFactories. In these factories, altogether 327,053 workers were employed. Within the 6 months before October 2008, 205 factories were visited unannounced by BetterFactories (21st Synthesis Report of BFC). 27 international buyers are taking part in the program so far, all of them of either US or European origin. The work of BetterFactories is specialized in the monitoring and reporting of working conditions in Cambodian garment factories according to national and international standards. BetterFactories supports the factories in improving the conditions and their productivity, and to ensure transparent improvement. The program is mandated and authorized by international buyers to improve the conditions of the garment workers of the supplying factories, which is part of their CSR strategy.

Trained BFC personnel carry out visits to each factory every 6 months in order to check the changes and improvements. These visits are said to be unannounced, and their checklist comprises about 500 items. Methods include interviews with managers as well as confidential on-site interviews with the workers, union members and shop stewards. They also include an inspection of the factory, the worker’s accommodations, and the area around the factory. After each visit, a detailed report is being provided in which the results and changes are compiled and recommendations are given. A few months after the main visits, each factory is visited again in order to investigate the changes that have taken place. Besides this auditing, BetterFactories is providing a range of services to support the workers to improve the working conditions. These services include trainings and further education which cover issues like global connections in textile production, project management, labour standards as well as rights and health education. The cooperation of the factories and garment unions is also a focal point of the training programmes of BetterFactories.

The main reason for selecting BetterFactories as a second case study is the localisation of the CSR-activities inside the core business: garment workers are neither shareholders nor consumers, but evidently direct stakeholders. This offers the opportunity to conduct research within the area of built-in CSR. There is an existing connection between the companies and the garment workers that would also exist without the CSR-activity, hence, changes and impact can be observed directly in the factories and the lives of the garment workers. The constellation of the stakeholders is very interesting: the international buyers and the garment factories as protagonists of the CSR-involvement, the program of BetterFactories as intermediary instance, and the workers of the factories as target group.

Analysis of research results

For the purpose of analysing the output of BetterFactories, the main CSR-Input of the involvement will briefly be summarized here: BetterFactories conducts monitoring visits to the factories, BetterFactories provides trainings for workers and managers, the program publishes the results for the public and BetterFactories provides expert knowledge about the topic.

This drafted Input leads to an outcome that can be differentiated in eight different main issues: deficits and improvements are identified that are related to either the social working conditions or to the workplace working conditions, workers know more about issues either directly or indirectly related to their work, managers know more about issues regarding the employment of garment workers, the public awareness has increased (factory owners are aware of this), international buyers can ask for advice related to their work with the Cambodian garment factories, and most of all: the international buyers can read the results and decide whether they want to continue their contract with the particular garment factory. This puts pressure on the factories to improve the situation, whereas the public awareness puts pressure on the international buyers to only work with factories that comply sufficiently (Table 6).

Table 6 Overview of CSR-input and first outcome of the involvement of BFC
Full size table

Again, the last of the three steps contains the impact analysis of the program. Here, I will evaluate the impact on each capability and then again sort it into the developed matrix.

The impact of BetterFactories on income is comparably large. Not only has the factual income increased, but minimum wages are paid more often (97% instead of 13% of the factories now pay their workers the legal minimum wage of US $40 per month (21st synthesis report)) and the in-time and overtime payment has increased as well (17th synthesis report). Right now, almost all (97%) of the factories comply and pay the regular legal minimum wage (21st synthesis report), which, however, still does not cover the living expenses in most cases. There is less perspective that this income will further increase in the near future, unless the legal minimum wage rises again. Therefore the direct indicator can be confirmed, whereas the indirect indicator cannot. The impact is both visible in the short term as well as from a longer perspective. The income has increased very quickly after the first involvement of BetterFactories and continued to increase over the past several years.Footnote 6 It is clear that the impact can be attributed to the involvement of BetterFactories, because there are no other factors that could have influenced the income policies in the factories except an altruistic change which is very unlikely according to the interviewed experts.

The impact of BetterFactories on the individual level of education is strong from a longer-term perspective, but also visible in the short term. Regular schooling possibilities amongst garment workers are rare and the perspective to be better educated is immediately increased after the involvement of the program. However, sustainable effects can only be seen at a later time. There are training courses offered by BetterFactories in which both the workers and the managers can take part. There are some training programmes offered by the factories which go beyond job training and include the teaching of English and accounting. Without a doubt, the establishing of these training possibilities as well as childcare can be attributed to the involvement of BetterFactories. It can be stated, that the average knowledge of workers in the garment factories has vaguely increased (measured in their knowledge of English for example). The direct indicator can therefore be confirmed. The indirect indicator cannot, because there is no perspective, that there will be more courses offered in the future.

The impact of BetterFactories on women’s empowerment is comparably large. There is evidence, that there is less sexual harassment and discrimination against women within the factories and income gaps between men and women have narrowed (Makin 2006:i). Women have started to set up female garment unions to fight for equal rights. This can partly be attributed to the involvement of BetterFactories and partly to the general upcoming of garment unions in Cambodia. BetterFactories provides a training program for women empowerment. The impact on youths however is comparably low. Young textile workers have problems accessing the factories, because the employers fear fake IDs of workers that conceal their real age. Before the involvement of the program, there has been evidence of several cases of underage workers in the factories, whereas now, there is only one evident case of underage work. Anyhow, the numbers here are too small to be significant. The impact is strong both on the long-term as well as the short-term perspective. Sexual harassment has decreased immediately after the first monitoring visits, according to the monitoring results, which, however, do not always portray the reality. Both indicators can therefore be confirmed. It is not directly clear if the impact can be attributed back to the involvement of BetterFactories, because there might be other factors which majorly influence the empowerment of women, for instance, a potential general emancipation of women in Cambodia.

The impact on the level of individual health is visible both immediately and in long-term perspective since there are immediate changes that are brought along by BetterFactories (the implementation of a medical staff for instance) and these changes last for longer. There are less diseases and illnesses amongst the group of workersFootnote 7 and there are health education programmes provided which go along with the regular education programmes. Therefore, both indicators can be confirmed. Furthermore, there are now regular medical check-ups in the factories.Footnote 8 It is clear that the impact is rooted in the involvement of BetterFactories, since there were no other significant factors to influence the average individual health during the last decade.

The impact on the capability of an adequate accommodation can only be seen from a longer perspective. Most of the workers still live in barracks around the factories. There have been very few cases in which the increased income was enough to rent a more adequate accommodation. However, this number is too small to be of any significance, and it can be attributed back to BetterFactories, since the program is mostly responsible for the income increase of the workers. However, the effect is only a small side effect to the income increase. None of the indicators can be confirmed. There is also a slight chance that due to increased income the indirect indicator can be confirmed in the future.

The impact on labour conditions is most likely to be the most significant impact of BetterFactories. It is evident that in many cases, labour conditions have improvedFootnote 9 and there are programmes and trainings for the workers to discuss and learn about their working conditions. Both indicators can therefore be confirmed. This improvement of labour conditions can be attributed back to better labour laws, widened media attention, and therefore the interest of the big companies in it. BetterFactories as part of this involvement is responsible for part of this improvement in labour conditions. The impact on labour conditions is visible immediately after the beginning of the involvement, but even more in longer terms. The impact can be attributed to the involvement of BetterFactories, since there have been direct recommendations of the program towards labour conditions which have been implemented by the factories.

There is almost no impact on the access to capital, land and insurance. The workers do not regularly get loans, purchase land, or are insured, nor has the income increase so far made any of that possible. The latter is the only slight perspective on this capability, since there is a perspective of higher income in the future.

The impact of BetterFactories on the access to the regular health care system is evident. Although, almost none of the workers have proper health insurances, more workers do than before. Due to the medical staff on duty during working hours and overtime, regular health check-ups are possible and therefore improve the workers’ access to the healthcare system related to their work place. This improvement has taken place immediately after the involvement and will possibly last sustainably. Since 2004/05 the factories pay visits to hospitals and doctors, if the worker can confirm that the illness was caused by or during their work time. This concession to the workers access to the health system was induced by massive pressure from BetterFactories. This does not, however, confirm the indirect indicator for the access to health system, since it does not cover regular hospital visits that are not related to work. The impact regarding the access to health services can partly be attributed to the involvement, since the medical staff on duty has been one reason for the employee’s regular check-ups. The direct indicator can therefore be confirmed. However, the general change in the access to the health care system cannot be attributed to BetterFactories since the program is only involved in the improvement on the work place.

Both indicators regarding the access to social networks can be confirmed in this case study. The involvement has both created an easily accessible social network as well as a possibility for the target group to express its joined opinion together. There are many more strikes than a decade ago, which are, however, still heavily (and sometimes violently) suppressed by many of the factory managers. The workers tend to begin to stand up for their labour rights and make a personal effort to support the whole group which is a relatively new phenomenon amongst garment workers. The impact is more easily visible from a long-term perspective, because the creation of the “worker’s identity” took a certain amount of time. The change in the existence of this capability can be attributed back to various factorsFootnote 10 of which BetterFactories is most likely to be one. Hence, the impact is difficult to attribute back to the program.

The direct one of the two indicators can be confirmed regarding the capability of social security and access to legal system: there is less criminal violence towards the individual and less danger at the work place and there is evidence, that most factory owners have improved the security standard at the factories (21st synthesis report). Furthermore because of the newly established transport methods from Phnom Penh city to the factories in the outskirts, the danger of traffic accidents, one of the major security threats, has slightly decreased. The direct indicator can therefore be confirmed. But due to the preventive nature of the measures, a potential impact is difficult to trace back to BetterFactories. However, there have been no changes in the way that workers can enforce their labour rights on a legal basis. Workers fear violent repression of both employers and police. Frequently, the police interrupt strikes and demonstrations of workers where many of them have been injured. In 2004, there have been four assassinations of union leaders. Taking part in unions and especially being active in them is still very dangerous. Due to international media coverage, the legal issues of garment workers have been put into society’s interest. However, this cannot be attributed to BetterFactories (Tables 7 and 8).

Table 7 Classification of the capabilities into the analysis scale regarding BetterFactories
Full size table
Table 8 Ranking of impact of BetterFactories regarding different capabilities
Full size table

Conclusions

By attempting to draw conclusions from the collected data in this article, I am fully aware of the fact that “reaching perfectly certain conclusions from uncertain data is obviously impossible. Indeed, uncertainty is a central aspect of all research and all knowledge about the world” (King et al. 1994: 8f.). By measuring the impact of CSR-strategies on Sen’s notion of poverty, some substantial methodological challenges had to be faced, due to the complex nature and the different understandings of the phenomenon CSR (cf. Jonker and Marberg 2007: 108). The main challenge was to find out which effects on poverty were induced by CSR and which were not. I faced this by going into micro-level detail regarding the influential mechanisms on every capability, and conduct a process tracing in order to isolate other factors that affect poverty.

At the first glance, the case studies have revealed that CSR influences almost all of the considered capabilities. However, at a second glance, some further implications can be drawn: firstly, the influence varies tremendously throughout the cases. Whereas for instance the impact on labour conditions in the BFC-case is somehow obvious but diverse, the impact on the same capability in the CS-case is rather mediocre. Additionally, the impact on poverty also varies immensely depending on the capability. It becomes clear that certain capabilities are more affected than others. For example, the impact of both CSR-strategies on the access to capital and on accommodation is rather small, whereas the impact of both CSR-strategies on the level of education or the access to social networks exceeds the former considerably. Therefore, it can be inferred that certain capabilities are more accessible to external influences than others. Despite these variations, there are some remarkable differences between the impact of the two types of CSR to be noted (see Table 9).

Table 9 Differentiation of impact of BetterFactories and ChildSafe regarding different capabilities and the different routing questions
Full size table

It became apparent that BetterFactories tends to have slightly more impact on the capabilities than ChildSafe, since the impact of BetterFactories on almost every capability scores higher than ChildSafe, except with regard to security, access to social networks and level of education.

Even if the results might only reflect a slice of reality, I will now turn to reconsider the hypotheses. The research results, thereby, might facilitate some interesting insights into causal mechanisms regarding CSR and poverty reduction:

First hypothesis: does CSR really have an impact on poverty as the first hypothesis assumes and as stated by Newell and Frynas (2007), Wolff et al. (2009), Volkert (2009) and many others? Does the private sector really have the potential to reduce poverty (Nelson 2006: 41 and 47)? The case studies revealed that there is a certain impact on poverty by CSR measures with regard to almost every capability and this impact can (at least partly) be attributed to the private sector involvement. It can therefore be stated that both the implementation of built-in and bolt-on CSR-strategies can have positive effects on the lives of the stakeholders. However, this is a simplification of reality, since the “wide variation in the specific contributions that (…) companies can make to supporting (…) poverty alleviation” (Nelson 2006: 46) could be empirically revealed. One of these variations is the difference of the impact between both kinds of CSR-strategies.

Second hypothesis: is the impact of built-in CSR more effective than the impact of bolt-on CSR as assumed in the second hypothesis? Following this, the BFC-case would score higher in regard to poverty reduction than the CS-case. This assumption proved to be true and the argument that “the greatest (…) contribution that any company can make to development is through carrying out its core business activities in a (…) productive and responsible manner” (Nelson 2006: 47) seems to be correct. However, I truly disagree with Newell and Frynas (2007) who argued that 99.9% of the impact is through core business activity, since the bolt-on CSR of ChildSafe did have a considerable effect on poverty.

As stated above, Wolff argues that “sustainability is impaired, when companies do not regard the […] CSR issue as strategically important” (Wolff et al. 2009: 295f.). On the ground of my research, I partly agree with this argument. It is obvious that companies do tend to put more efforts in CSR-activities that relate to their interests. It could be seen in the BFC-case that garment factories improved the situation for workers especially related to issues that were visible to monitoring and international buyers, e.g. the payment of wages or safety measures at the work place. Other capabilities, which were not as obvious to the program, were improved to a lesser extent. The access to capital, accommodation, or the establishment of social networks are much less controllable by the monitoring staff and therefore the garment factories seemed to put less effort into the improvement here. Barrientos and Smith (2007) confirm this argument and state that what they call ‘outcome standards’ (e.g. health and safety issues) are more often positively influenced by CSR measures than the so-called process rights (e.g. freedom of association). Furthermore, the hypothesis says that business that apply built-in CSR have a stronger motivation to assess their impact thoroughly. The research confirmed that the reasons for companies to implement CSR measures vary. In the BFC-case, the garment factories as well as the international buyers demonstrate their responsibility towards their stakeholders in order to satisfy consumer expectations. In the CS-case, the small corporations demonstrate responsibility towards street children in order to please their customers, with the intention to enhance the consumption of their products.

So although, it is widely claimed that bolt-on CSR is not as effective as built-in CSR, it did have quite large effects in the regarded ChildSafe case. This could be attributed to the cooperation of the businesses with a renowned organization, which balances the negative effects of companies only conducting philanthropic measures themselves. Lepoutre and Heene also argue in line with this and state that “shared responsibility and the creation of institutions for joint responsibility taking” (Lepoutre and Heene 2006: 267) increases the impact of the CSR-activities of very small corporations (cf. Banerjee 2008).

Thus, it can be concluded that CSR does have a certain impact on poverty reduction and the correlation between corporate social responsibility and the reduction of poverty in the light of Sen’s approach has become somewhat apparent throughout both case studies. However, the matter is slightly more complex than this. A wide range of context-dependent factors influence the impact of CSR of which I addressed one in this paper, namely different forms of CSR. Still, there are far more implications to be considered in further research regarding the relationship between corporations and the affected stakeholders, and the debate on the scope and the complex effects of CSR has to be continued.