Beyond shampoo, salt and SIM-cards : an analysis of BOP initiatives reported by the Global Fortune 250.

(1)

Beyond Shampoo, Salt and SIM-cards

An analysis of BOP initiatives reported by the Global Fortune 250

Master Thesis by Khadija van der Straaten 1st Supervisor: Prof. dr. Ans Kolk Student number: 10476024 2nd Supervisor: Dr. Arno Kourula MSc Management Studies, Strategy Track Submission Date: February 13, 2014

(2)

Acknowledgments

First of all I would like to thank Dr. Arno Kourula for introducing me to the topic of this thesis during his inspirational lectures on business and sustainability. Furthermore I would like to thank Prof. dr. Ans Kolk for her valuable feedback and support and Dr. ir. Dick Swenne for proofreading the final draft.

(4)

Abstract

Prahalad and Hart’s (2002) proposition that multinational companies can alleviate poverty by profit generating activities at the bottom of the world’s income pyramid has received much attention in the business literature. Most of the studies on the topic however are restricted to a limited set of companies and countries, and a considerable number have local companies and NGOs as a unit of analysis. The question arises whether the cases studied in the literature are representative of the actual business activity of large (multinational) companies. The large-N analysis of bottom of the pyramid business initiatives vis-à-vis the cases studied in literature needed to answer this question has been made in this study. The self-reported initiatives of the Global Fortune 250 (GF250) and the cases studied in recent literature were systemically analyzed and compared, in order to reveal their similarities and differences. 79 of the GF250 companies reported bottom of the pyramid initiatives, and the characteristics of both these companies and their initiatives were found to be different from the characteristics of the cases studied in the literature in several ways. The most notable difference was the very low level of attention paid to non-Western multinational bottom of the pyramid initiatives in literature compared to the number and scope of these activities reported by the GF250 themselves. This has led to a number of recommendations for improvements in both research and GF250 reporting on the bottom of the pyramid concept.

(5)

1. Introduction

“The unorganized but the indispensable units of economic power, that build from the bottom up and not from the top down, […] the forgotten men at the bottom of the economic pyramid.”

(Franklin D. Roosevelt)

Banerjee and Duflo (2007) describe the pressing problems of the people living at the bottom of the world’s income pyramid; how they have to go without a single bite to eat for an entire day, lack clean drinking water and sanitation with ailment, unhappiness and weakness as a result. Since the term was introduced by Franklin D. Roosevelt in 1932, people at the top of the pyramid have increasingly taken an interest in the problems of the people at the bottom of the pyramid, by making generous donations to non-governmental organizations (NGOs) and electing governments which address the needs of the poor for example. Although a lot of progress has been made by these institutions, extreme poverty persists, however, with almost one billion people expected to still be living below the US$1.25 a day poverty line in 2015 (United Nations, 2012).

Considering this failure of governments and NGOs to eliminate poverty it was no surprise that Prahalad and Hart’s (2002) proposition that multinational companies can alleviate poverty by profit generating activities drew much attention. According to Prahalad and Hart, multinationals could make a vast amount of profit at the lowest-income markets and alleviate poverty simultaneously if only they would overcome their Western orthodoxies. Since then, their bottom of the pyramid proposition has been investigated in over 300 business research publications. However, despite its origin in the international business strategy field, bottom of the pyramid literature seems to have moved away from the multinational company as the unit of analysis (Kolk, Rivera-Santos & Rufín, 2012). Literature published since Prahalad and Hart´s (2002) first publication, such as Akula (2008) and London (2009), focusses on local initiatives and NGOs as the source of poverty reduction or alleviation, thereby relaxing the original bottom of the pyramid proposition which states that poverty

(6)

can be alleviated by the profit generating activities of multinational companies in particular (Prahalad & Hart, 2002).

Furthermore the vast majority of bottom of the pyramid publications is based on conceptual examples or case studies, and these examples and cases are mostly from India. The set of companies used as examples or case studies is also quite limited and repetitive and Prahalad and Hart’s (2002) original example of Unilever India (HLL) is still the most cited company in the bottom of the pyramid literature by far. A little more variety exists in terms of the industries studied, with information and communication technologies, household products (such as detergent, shampoo and salt) and financial services being the main areas of research (Kolk et al., 2012).

The question arises whether the literature adequately reflects the actual business activities of large (multinational) companies. Are private (multinational) companies not as active at the bottom of the pyramid as predicted by Prahalad and Hart’s (2002) proposition? And are these activities restricted only to the countries and industries bottom of the pyramid publications seem to revolve around? To answer these questions, the self-reported data of the Global Fortune 250 (GF250) was systematically analyzed, in order to reveal the characteristics of large companies active at the bottom of the pyramid, and the characteristics of their activities. Recent (2010 and onwards) bottom of the pyramid literature was simultaneously analyzed to reveal the similarities and differences between bottom of the pyramid initiatives reported by the GF250 and business scholars, thereby answering this study’s main research question: “How do the bottom of the pyramid initiatives reported by the Global Fortune 250 compare to the initiatives studied in literature?”

This study aims to make a contribution in three ways. First, by extending Kolk et al.’s (2012) systematic literature review covering the years 1999 – 2009 by recent literature, this study includes an assessment of the most recent trends in bottom of the pyramid research, an assessment that has not yet been made. Second, this study is the first to include all of the 250 largest companies in the world, and will thereby be the first to examine their self-reported bottom of the pyramid activities. A

(7)

large sample examination of bottom of the pyramid activities is a more rigorous way to show the relevance of the bottom of the pyramid concept to business practice than the conceptual and case studies that are commonly used in bottom of the pyramid research. Third, the empirical analysis of the similarities and differences between bottom of the pyramid initiatives reported by the GF250 and the business literature is an approach to identify recommendations for further research that has not been used before within bottom of the pyramid research.

In the next section of this thesis, the definition of the bottom of the pyramid and the theoretical foundations of the bottom of the pyramid proposition will be presented in order to set the scene for the subsequent sections discussing the study itself. Because an abductive research approach was used, the methodology and conceptual framework were developed simultaneously; the resulting research design will be drawn in section three. Next, the empirical results of the literature review and GF250 data collection are presented in the fourth section of this thesis, followed by the discussion and conclusions section.

(8)

2. The Bottom of the Pyramid Proposition

In this section, the definition and characteristics of the bottom of the pyramid will first be presented. Subsequently, the two key elements of the bottom of the pyramid proposition will be discussed: poverty alleviation and profit generation resulting from (multinational) business activity.

2.1 Bottom of the Pyramid definition

A significant amount of debate exists on the definition of the bottom of the pyramid itself. In their initial (2002) article, Prahalad and Hart define the bottom of the pyramid as the 4 billion people living of an income under $1500 a year at purchasing power parity. Karnani (2007), on the other hand, argues that the number of people in that income class is smaller, and purchasing power parity is irrelevant from a multinational companies’ point of view. Prahalad (in London & Hart, 2010) responds to this criticism by pointing out that the number of people living at the bottom of the pyramid, and whether their income is calculated at purchasing power parity or not, is inconsequential to the validity of his proposition. London and Hart (2010) agree that the discussion on the exact calculation of the income and the number of the poor is distracting attention away from the value of the bottom of the pyramid proposition and should be left to the World Bank, not to business scholars. They propose defining the bottom of the pyramid by a broader set of characteristics that are relevant from a managerial point of view.

The first of these characteristics of the bottom of the pyramid is exclusion from the formal capitalist economy (London & Hart, 2010; Martinez & Carbonell, 2007). People at the bottom of the pyramid usually do not hold title to their assets, and are therefore excluded from sublending their assets for formal loans (Banerjee & Duflo, 2007). Furthermore, they usually lack access to the financial system altogether (e.g. they don’t have bank or savings accounts) and engage in inefficient transactions with local informal enterprises (Rangan, Chu & Petkoski, 2011). Microfinance is designed to solve some of these issues; its outreach is still limited, however (Khawari, 2004).

(9)

Apart from the lack of access to finance, people at the bottom of the pyramid lack access to complete market information. This is likely a result of their low level of education and skills, simultaneously assessing products in terms of price and quantity, for example, can be challenging for the poor. This challenge is aggravated by exposure to the media and results in an even greater dependence on local informal enterpises (Tarafdar, Anekal & Singh, 2012; Ramani, SadreGhazi & Duysters, 2012; Viswanathan, Gajendiran, & Venkatesan, 2008). These circumstances lead bottom of the pyramid consumers to make decisions that are irrational from an economical point of view, to a higher degree than top of the pyramid consumers (Ramani et al., 2012). For companies active in the bottom of the pyramid, this makes it hard to estimate demand, which prevents transactions with a positive consumer and producer surplus from taking place (Tarafdar et al., 2012; Viswanathan & Rosa, 2007).

The lack of physical access is another challenge for both consumers and companies at the bottom of the pyramid market. A large number of bottom of the pyramid consumers live dispersed and far away from the sites where formal business transactions take place, and lack adequate infrastructure. Many of the bottom of the pyramid business models are aimed at overcoming these challenges (Prahalad & Hart, 2002; Tarafdar et al., 2012).

The second of London and Hart’s (2010) bottom of the pyramid characteristics is its heterogeneous nature (London & Hart, 2010). Research by Banerjee and Duflo (2007) shows a wide variety in, for instance, living conditions, access to health care and education, and possession of assets among the poor within and between countries. Furthermore, they found differences between the characteristics of the people earning an income less than $1 per day and more than $1 per day within countries. Due to this heterogeneity, Prahalad (2010) even states that “there is no universal definition of the bottom of the pyramid that can be useful” (p. 7) because the target group of a bottom of the pyramid initiative must be defined by its specific characteristics in order for the initiative to fit the needs of the target group.

(10)

Rangan et al. (2011) segment the bottom of the pyramid market in two ways, by income and by value creation role. In terms of income they find, similar to Banerjee and Duflo (2007), that people with an income of less than $1 a day, the extremely poor, lack basic needs, make very irregular or barter transactions, and spend most of their time trying to survive. People with an income between $1 and $3 a day, living at subsistence level, usually earn irregular incomes as temporary workers, and have only access to poor quality basic needs and the informal market. Those earning between $3 and $5 a day, termed the low-income segment, usually have semi-permanent employment, and (limited) access to basic needs and the formal and informal market, and possess some assets.

In addition to understanding the different income segments, Rangan et al. (2011) stress the importance of segmenting the target market by value-creation role for businesses operating at the bottom of the pyramid. The first segment, clients, consists of beneficiaries of aid and assistance that are usually located in the extreme poverty income segment. The second segment, (co)producers, usually consists of people in the subsistence income segment, and their main barrier to escaping poverty is finding permanent employment. Finally, the third segment, consumers, usually consists of people in the low-income segment, primarily because they are the only ones at the bottom of the pyramid who can afford to consume.

In their final set of bottom of the pyramid characteristics, London and Hart (2010) do attempt to define the bottom of the pyramid in quantitative terms, be it in a very broad manner. They state that the bottom of the pyramid is composed of the people with the least amount of income, who account for the majority of the world’s population. While these characteristics should be present for the bottom of the (income) pyramid to actually be the bottom of the pyramid, they offer little additional insight for business and academic purposes.

2.2 Poverty alleviation

The first key element of the bottom of the pyramid proposition is whether, and if so, under what circumstances, profit generating business activities alleviate poverty.

(11)

In bottom of the pyramid research the construct poverty alleviation is usually poorly defined, which poses a problem to the rigor of the concept (Kolk et al., 2012). Prahalad (2010) and Martin and Hill’s (2012) examples, for instance, include cases in which poverty alleviation takes the form of additional product offerings or sources of empowerment to bottom of the pyramid consumers, with no rigorous explanation as to whether and how this could actually reduce poverty (Landrum, 2007). According to Karnani (2007), only poverty reduction is relevant, because the poor are only better off when bottom of the pyramid initiatives reduce prices or increase income. Because the poor are often constricted by lack of resources, not by lack of product offering, tempting the poor to reallocate resources away from their basic needs might even be considered exploitation, not poverty alleviation (Hammond & Prahalad, 2004; Rangan et al., 2011). Selling to the poor might even have an adverse effect on poverty when it hurts local business (Warnholz, 2007).

London, Anupindi and Sheth (2010) focus on Rangan et al.’s (2011) producer segment as the location where both profits and poverty alleviation, termed mutual value creation, can simultaneously be achieved. They identify the enhancement of resources such as local networks as the source which creates value for the poor. Still it remains questionable whether and how this would increase their income or reduce prices and thereby reduce poverty.

Furthermore, bottom of the pyramid initiatives aimed at Rangan et al.’s (2011) consumer segment could potentially be a location for poverty reduction as well, as described by development economists Melamed, Hartwig and Grant (2011) and Loayza and Raddatz (2009). Their work offers insights into the underlying economic mechanisms which might explain how business initiatives at the bottom of the pyramid may reduce poverty. Establishing how multinational activity can positively affect economic development (e.g. Clay, 2005; Fortanier & Kolk, 2007) is insufficient in this respect, as economic growth on a macro level does not necessarily benefit the poorest part of a country’s population (Kakwani & Pernia, 2000). According to Melamed et al. (2011) and Loayza and Raddatz (2009) several conditions have to be met in order for poverty to be reduced:

(12)

• Multinational activity should introduce economic growth, which in turn should increase demand for labor (i.e. “jobless growth” must not be present);

• This demand should be for low-skilled labor (increased demand for high-skilled labor will increase income inequality);

• Labor productivity should increase simultaneously; considering that 40% of workers worldwide earn an income below the US$2 a day poverty line, an increase of productivity is needed in order to increase wages;

• Gender inequality and youth unemployment have to be addressed, as these issues affect growth negatively.

From a business point of view understanding these mechanisms is relevant to avoid being accused of exploiting the poor, which poses a risk to the company’s reputation and image, and because reducing poverty enhances the potential for scalable profits (London & Hart, 2010).

An entirely new point of view of the dynamics of poverty has recently been introduced by Mullainathan and Shafir (2013). Their research on the psychology of scarcity shows that when people experience a pressing sense of scarcity (especially scarcity of money) their intellectual capacity to make sound decisions decreases. Poor decision-making (e.g. spending when the person should be saving or spending on luxury goods) reinforces poverty and thereby the sense of scarcity. According to Mullainathan and Shafir, facilitating easy decision-making (e.g. presenting information in one dimension at a time) results in better decisions. Moreover, this frees up intellectual capacity to make sound unfacilitated decisions; this improved decision-making can in turn reduce scarcity of money and consequently increase a person’s intellectual capacity, which breaks the reinforcing cycle of poverty. Their research could be the “missing link” between poverty alleviation and poverty reduction, as poverty alleviation may reduce the sense of scarcity, thereby increasing intellectual capacity and sound decision-making, and eventually reducing the actual level of poverty.

(13)

2.3 Profit generation

The second, and strategically most relevant, element of the bottom of the pyramid proposition is whether profit can be made at the bottom of the pyramid.

According to Karnani (2007), the dispersion of the people living at the bottom of the pyramid, combined with poor infrastructure, results in high distribution costs. Moreover, the heterogeneous nature of the bottom of the pyramid prevents opportunities for economies of scale. These high costs combined with the low purchasing power of the poor leave no room to make a profit in the bottom of the pyramid market, Karnani (2007) concludes.

Although a clear survivor bias is present in their case selection, numerous authors including Prahalad (2010) and Seelos and Meir (2007) have since described business initiatives which were successful in making a profit in bottom of the pyramid markets, which demonstrates that an opportunity to make a profit does exist. Due to the vast majority of bottom of the pyramid research being based on either a limited number of case studies or conceptual examples (Kolk et al., 2012), no conclusion can be drawn about the overall level of profitability and success in the bottom of the pyramid market; for every example of a successful company there could be ten that failed, or not. Furthermore, the number of successful companies and the size of their profits are unknown, making an assessment of the actual size of the potential fortune at the bottom of the pyramid impossible.

A number of bottom of the pyramid studies focus on external conditions as key determinants for profitability, such as the presence or absence of a complementary infrastructure as discussed by Ramani et al. (2012) and Karamchandani, Kubzansky and Lalwani (2011).

However, the (elements of the) company’s business model which can lead to profitability at the bottom of the pyramid has been the focus of the majority of bottom of the pyramid studies within the strategy field. Pitta, Guesalaga and Marshall (2008), for example, focus on marketing as a key determinant of bottom of the pyramid success, as do Weidner, Rosa and Viswanathan (2010), while

(14)

Hart and Christensen (2002) and Simanis and Hart (2009) focus on innovation as the main determinant of profitability.

Apart from apparent disagreement on the determinants of profitability, there is divergence between the conclusions made on the same determinants as well. Akula (2008) for example stresses the importance of high volumes at low margins of being profitable at the bottom of the pyramid, while Simanis (2012) argues that companies need to set high margins and price points.

Given the importance attributed to business models in the literature, the type of business model used was the starting point of the abductive process of data collection and analysis for this study, as will be discussed in the next section.

(15)

3. Research Design

Considering that few studies have been done on the characteristics of companies active at the bottom of the pyramid, and a wide variety of characteristics of their initiatives have been studied but rarely in a systematic manner, this research design was developed in an abductive manner. This implies that the research design and data analysis were both based on an iterative bidirectional approach; the research design was adapted as a result of matching existing literature with the emerging concepts from the data, and the data analysis was redirected by the adapted framework (Dubois & Gadde, 2002; Shepard & Sutcliffe, 2011).

3.1 Literature review

A systemic review of bottom of pyramid literature was conducted over the period January 2010 up to and including August 2013; the methodology was based on the one used by Kolk et al. (2012). First the abstracts and citations of peer-reviewed articles where searched by entering the keyword “bottom of pyramid” in Ebcohost. This action returned 152 results, of which 37 were not related to bottom of the pyramid (e.g. physics-related research on nanopyramids) and therefore excluded. The search was repeated using the keyword “base of pyramid”, returning 234 results. Of these results, 154 were not related to bottom of the pyramid and 19 were exact duplicates of “bottom of pyramid” hits and therefore excluded. The exact same keyword search was then entered in Proquest returning 29 additional results for “bottom of pyramid” and 18 additional results for “base of pyramid”. Of the “bottom of pyramid” results, 7 were not related to bottom of the pyramid and of the “base of pyramid” results, 12 were not, all of which were excluded, as was the one duplicate resulting from the two Proquest searches. The keyword search was then repeated for articles published in the Harvard Business Review, which resulted in 5 additional hits.

In total 208 unique publications on the bottom of the pyramid were thereby identified. Subsequently, the content of the 208 publications was examined. 5 publications were not available via standard sources (e.g. an Albanian conference proceeding and an erratum). 71 publications were

(16)

discarded because they did not make new research contributions (literature reviews, book reports, etc.) or had microfinance as their main topic. The publications on microfinance were excluded because according to Kolk et al. (2012) they belong to a “distinctly different scholarly stream” (p. 2).

The 132 remaining publications were then coded by unit of analysis and academic field in order to identify which articles were relevant for comparison vis-a-vis the cases reported by the GF250; only the 59 strategy-related publications (listed in Appendix I) with a firm focus were qualified as relevant.

The relevant publications were then coded:The starting point of data collection was the organizing framework developed by Kolk et al. (2012); a framework that classifies bottom of the pyramid business initiatives by position of the poor in the value network and their mode of engagement. In terms of position of the poor in the value network, Kolk et al. (2012) distinguish between producers, consumers and employees. They use three categories of mode of engagement: coinventors of the business model, recipients of an adapted model and recipients of a non-adapted model. When coding the articles, a distinction between the modes of engagement was made based on the most commonly used framework of Osterwalder and Pigneur (2010).

(17)

Their framework, or business model canvas (figure 1), consists of nine building blocks, which together comprise a company’s business model. The first building block is the customer segment or segments that a company serves. The second building block is the value proposition; the products and/or services the company offers to meet the needs of the customer(s). The channels, the third building block, are the communication, distribution and sales channels the company uses to deliver its value proposition to its customers. The fourth building block defines how customer relationships are established and maintained (e.g. via dedicated personal assistance, self-service or automated service). The fifth and sixth building blocks are financial in nature: the company’s revenue streams and cost structure.

The key resources, the seventh building block, are the assets that are needed to be able to offer and deliver the value proposition to the customers, and the eighth building block, key activities, are the activities needed to do so. Some key activities and resources are sourced from partners (e.g. key suppliers) outside the company, the key partners comprise the ninth building block. Expanding into bottom of the pyramid markets naturally implies a change in customer segmentation, therefore this building block was excluded from the analysis. A change in any of the other building blocks initiated by the company in order to better fit the bottom of the pyramid market was considered an “adapted business model”. When the poor were actively engaged in shaping this change, it was consider “co-invention”. When the eight building blocks were not specifically adapted for the bottom of the pyramid market, the initiative was coded as a “non-adapted business model”.

• Methodology; whether the study was quantitative or qualitative, conceptual or empirical, and in case a case study was performed, whether this case study was a single or a multiple case study.

• Whether a multinational company was involved, and in case it was, which company it was. • The county (or countries) and NAICS sector (or sectors) the study was set in.

(18)

• Whether any and if so which partners (NGO or government) were involved in designing and executing the business model.

• The outcome of the case on all dimensions of the triple bottom line (Kolk et al., 2012); economic, social and environmental outcome of the business activity for the target group, firm and partners, if reported.

3.2 GF250 characteristics

No systematic studies have been done on the characteristics of the companies reporting on bottom of the pyramid initiatives. A number of studies however, have been done on the characteristics of companies reporting on their sustainability and corporate responsibility in general. During data collection, it became apparent that many companies reported on their bottom of the pyramid initiatives as part of their sustainability/corporate responsibility reporting. Moreover, Kourula & Halme, 2008 describe the bottom of the pyramid concept as a specific approach to corporate responsibility. The selection of independent variables to predict bottom of the pyramid reporting was therefore based on the literature on sustainability reporting.

Kolk (2010) focusses on industry and country of origin as key determinants of corporate responsibility reporting by the GF250 and these determinants have thus been examined in this study as well. She identifies companies from smokestack industries such as chemicals and pharmaceuticals, electronics and automotives as the most active companies in reporting on corporate responsibility, while companies from more service-oriented industries such as trade and retail are the least active, with the exception of the financial and insurance industry. Moreover, she finds that over 90% of the companies from European countries and 74% of the companies from Japan report on corporate responsibility, whereas the number of American and South Korean companies that report is relatively low.

As discussed in the first section of this report, the original bottom of the pyramid proposition predicted the bottom of the pyramid to be an attractive market for multinational companies

(19)

(Prahalad & Hart, 2002), but many publications have since relaxed this element of the proposition. Moreover, Kolk and Fortanier (2013) found a negative relationship between a company’s level of internationalization and likelihood of reporting on sustainability (although their study only included the environmental aspect of sustainability). Because of the divergence within the literature, it will be included in the analysis whether there is a difference in reporting on bottom of the pyramid initiatives between multinational and non-multinational companies.

A significant relationship between UN Global Compact membership and bottom of the pyramid reporting was observed in the data. Theoretical support was found in Whitehouse’s (2003) publication; she established UN global compact memberships as a manifestation of corporate citizenship and defined corporate citizenship as a result of external encouragement to behave responsibly, as opposed to corporate responsibility that is a result of external enforcement. Many elements of corporate responsibility reporting, such as waste treatment and greenhouse-gas emissions are externally enforced on many companies. Serving the bottom of the pyramid and reporting on this, however, is not required for any firm and can therefore be characterized as corporate citizenship. The positive relationship between UN Global Compact membership and bottom of the pyramid reporting might therefore be explained be the level of corporate citizenship.

Chih, Chih and Chen (2010) study which characteristics determine a company’s position on the Dow Jones Sustainability World Index, a position that is in part dependent on corporate responsibility reporting. They identify firm profitability and size as two key control variables in their study. Firm profitability is included because less profitable firms have fewer resources to spare for responsible activities. Firm size is included because “large firms are subject to closer scrutiny by the public, thereby raising the likelihood of such firms acting in more responsible ways” (Chih et al., 2010, p. 117). Firm profitability and size were therefore included as control variables.

(20)

Data was collected from the top 250 companies in the Global Fortune 500; this list includes the world’s largest companies in a variety of countries and industries (see Appendix II). The GF250 2013 list was retrieved from CNN Money on August 19, 2013.

For every top 250 company, the “country of origin” was defined as the country in which the corporate head office is located, as reported on the companies’ corporate website. If the company reported sales and assets outside its country of origin on its corporate website, it was classified as a “multinational” (all companies had both foreign sales and assets or neither of them). Company “industry”, as reported on the CNN Money website, was confirmed by the information on the corporate websites, and reclassified into NAICS sectors at a three digit level based on the sector description from the NAICS (2012) (see the table in Appendix III for the conversion table), as the lower-level classification made by CNN Money did not reveal industry level results. Company “size” was measured by the company’s total assets (Chih et al., 2010) as reported by CNN Money. Company “profitability” was measured by the company’s return on assets (Chih et al., 2010) as reported by CNN Money. The financial data reported by CNN Money was checked by comparing it to the financial reports of 25 randomly selected companies; the data was correct in all 25 cases. UN Global Compact membership (“UNGC member”) was established via the participant search option on the UN Global Compact website. Only members with an active status were classified as members.

Subsequently, all corporate websites (English version only), most recent sustainability reports and reports submitted to the UN Global Compact were searched for initiatives aimed at the following target groups; “bottom of pyramid”, “base of pyramid”, “poverty/poor” and “low-income”, as these are the general terms used in literature to describe the bottom of the pyramid (Kolk et al., 2012). Any initiative (product, service or project) described to be aimed at these target groups was then counted towards the dependent variables: “BOP initiatives” (1 = reported, 0 = not reported) and number of BOP initiatives reported. Reports dated before 2010 were not included, neither were initiatives

(21)

reported as philanthropy, donations or employee voluntarism only. A total of 147 initiatives were identified.

The data was analyzed using the Statistical software Package for Social Sciences (SPSS). A linear regression model was constructed to test the hypothesized relations and control variables. First “BOP initiatives” was selected as the dependent variable. In step 1, the control variables total assets and return on assets were then entered into the equation. In step 2, the independent variables (multinational activity and UN Global Compact Membership) were entered into the model. Subsequently, the number of BOP Initiatives was selected as the dependent variable and the steps described above were repeated.

3.3 GF250 initiatives

The 147 bottom of the pyramid initiatives identified as described in the previous paragraph (see Appendix IV for the full list of sources) were coded by:

• The term used to identify target group (poor, low-income, base or bottom of pyramid) and, if reported, the income level of this target group at 2013 US$ purchasing power parity. The PPP conversion factor (GDP) to market exchange rate ratios published by the World Bank (2013) were used for all purchasing power conversions;

• The country (or countries) hosting the bottom of the pyramid initiatives;

• The economic, social and environmental outcome of the business activity for the target group, firm and partners, if reported;

• The bottom of the pyramid business model, in the exact same manner as the business initiatives described in the literature were coded. However, when coding the GF250 initiatives, a new category emerged: initiatives that provided products or services to the poor for free; this category was named “recipients”. This mode of engagement was added as a category to Kolk et al.’s (2012) framework; the resulting organizing framework is displayed in figure 2.

(22)

Figure 2: Organizing Framework (adapted from “Reviewing a decade of research on the “Base/Bottom of the Pyramid” (BOP) concept” by Kolk et al., 2012)

Consumers Co-inventors

Adapted BM

Recipients

Producers

Position in value network

M ode o f E ng ag em en t Non-adapted BM Employees

(23)

4. Results and Analysis

4.1 Literature review

After the article search and selection, the first thing that stood out was the number of bottom of the pyramid articles published since 2009. Figure 3 shows the number of published articles reported by Kolk et al. (2012) with the addition of the years 2010, 2011, 2012 and 2013 up to and including August, using the same inclusion criteria they used. The increase in the number of published articles from 2007 onwards has persisted, with a remarkable doubling of the amount of publications in 2012.

Figure 3: Number of Published BOP Articles Framework (adapted from “Reviewing a decade of research on the “Base/Bottom of the Pyramid” (BOP) concept” by Kolk et al., 2012)

A second observation was to the one noted by Schuster and Holtbrügge in their 2012 publication; the academic discussion on the bottom of the pyramid has moved focus from defining and estimating the market to examining specific aspects of the bottom of the pyramid market. The definition of the bottom of the pyramid in the most recent literature on the topic does not go further than Prahalad’s “4 billion” (e.g. Chikweche & Fletcher, 2012; Golja & Požega, 2012 and Rangan et al., 2011) or simply “the poor” (e.g. Anderson, Markides & Kupp, 2010; Karamchandani et al., 2011 and Ramani, et al., 2012) or “low-income” (e.g. Schuster & Holtbrügge, 2012 and Sesan, Raman, Clifford & Forbes, 2012).

(24)

The specific aspects of the bottom of the pyramid market considered in the most recent publications ranged from:

• Innovation for and by the bottom of the pyramid (e.g. Chakravarthy & Coughlan, 2011; Linna, 2012; Prahalad, 2012; Ray & Ray, 2010; and Ray & Ray, 2011) to

• Distribution and sales channels (e.g. Azmat, 2013; Dolan, 2012; Dolan & Johnstone-Louis, 2011; Dolan, Johnstone-Louis & Scott, 2012; Garrette & Karnani, 2010; Wentzel, Yadavalli & Sundar, 2013) to

• Marketing (e.g. Chikweche & Fletcher, 2012; Errington et al., 2012; Rahman et al., 2013 and Weidner et al., 2010) to

• Bottom of the pyramid franchise models (e.g. Alur & Schoormans, 2011; Chikweche & Fletcher, 2011 and Kistruck, Webb, Sutter & Ireland, 2011) to

• Internal organizational constraints of companies active in bottom of the pyramid markets (Halme, Lindeman & Linna, 2012).

Bottom of the pyramid research has moved on from mostly conceptual studies (Kolk et al., 2012) to a vast majority of more rigorous case studies. 26 of the 59 studies with a firm focus used a single case study; Hens’ (2012) studied Philips Lighting in Ghana and Singh and Srivastava’s (2012) studied Tata in India, for example. Another 26 studies were based on multiple cases; Schrader, Freimann and Seuring (2012), for example, made a cross-case analysis of the bottom of the pyramid activities of Allianz, BASF, BSH, Danone, Nestlé, P&G and Telenor. Only two studies used more than 10 cases in their analysis (London et al., 2010 (n=67) and Viswanathan & Sridharan, 2012 (n=13)).

In the selection of case studies, the “usual suspects” stood out. Prahalad and Hart’s (2002) conceptual example of Unilever was still the most used research subject in bottom of the pyramid research (e.g. Chakravarthy & Coughlan, 2011; Dolan, 2012; Dolan et al., 2012; Karamchandani et al., 2011; Rangan et al., 2011 and Schuster & Holtbrügge, 2012). Vodafone (e.g. Anderson et al., 2010; Schuster & Holtbrügge, 2012 and Weidner et al., 2010), Danone (e.g. Dolan et al., 2012; Garrette &

(25)

Karnani, 2010 and Schrader et al., 2012) and Hewlett-Packard (e.g. Rangan et al., 2011; Schwittay, 2011 and Schwittay, 2012) are examples of other well-known and widely used research subjects.

The majority of case studies studied the bottom of the pyramid initiatives of Western multinational companies in developing countries. Other case studies studied bottom of the pyramid initiatives by local companies. Ahuja (2012) and Tyagi (2012), for example, studied Jaipur Rugs in India, even though this case was also previously discussed by Prahalad (in his 2010 book “The fortune at the bottom of the pyramid”), as were the research subjects of Varman, Skålén and Belk (e-Choupal; 2012), Esposito, Kapoor and Goyal (Aravind Eye Care; 2012) and Simanis (CEMEX; 2012). Only one publication included a case study of a non-Western multinational initiative: Gardetti & D'Andrea’s (2010) study of Chilean timber company Masisa’s bottom of the pyramid activities in Argentina.

Figure 4: Geographic Origin of Cases in BOP Articles

The selection of countries is equally unsurprising (see figure 4). Out of the 41 studies that were performed in one or two countries, 20 were conducted in India (e.g. Ahuja, 2012; Esposito et al., 2012; Halme et al., 2012 and Ramani et al., 2012), again the setting of Prahalad and Hart’s (2002) original Unilver example. Other popular countries for bottom of the pyramid research are Bangladesh

(26)

(5 publications, e.g. Azmat, 2013 and Zainudeen, Samarajiva & Sivapragasam, 2011) and South Africa (4 publications, e.g. Dolan & Johnstone-Louis, 2011 and Wentzel et al., 2013). A notable exception is Errington et al.’s (2012) study of Nestlé’s marketing of instant noodles to bottom of the pyramid consumers in Papua New Guinea, a market not often studied.

41 of the 59 publications covered one or two industries. 21 of these revolved around 3 industries only: utilities (9 cases; e.g. Mahama’s (2012) study of the electricity market in Ghana), telecommunications (8 cases; e.g. Ilahiane and Sherry’s (2012) study of mobile phone use in Morocco) and information technology (5 cases; e.g. Schwittay’s (2011) study of the impact of Hewlett-Packard’s e-Inclusion program in Costa Rica).

In terms of modes of engaging the poor, very little has changed in the literature compared to the cases studied before 2010 (as described by Kolk et al., 2012). Over 90% of the articles described a company-initiated business model that was adapted for the bottom of the pyramid market to some degree, without involving the poor as co-inventors. Ramani et al. (2012), for example, studied how sellers of sanitation adapted both the design and distribution of toilets in order to reach bottom of the pyramid consumers, and Singh and Srivastava (2012) studied how the Tata Nano was able to gain market share by a marketing and distribution strategy designed specifically for the bottom of the pyramid.

Partnerships with NGOs were often described as a way to improve access to the bottom of the pyramid market (17 cases; e.g. Sesan et al.’s (2012) Clean-Cook case, in which the Swedish Dometic group uses an international NGO, Project Gaia, to explore the opportunities for its cooking stoves in bottom of the pyramid markets.) Governments were less often described as partners, though both partnerships with local and non-local governments were described. Halme et al. (2012), for example, described how Indian policy makers provided input to Nokia’s Village Connect developers, while Hens (2012) described how the Dutch government was closely involved in Philips Lighting’s design of products for the African bottom of the pyramid market. Despite the limited number of partnerships

(27)

described, institutional support was, however, often recognized as an important prerequisite for success in bottom of the pyramid markets (e.g. Hall, Matos, Sheehan & Silvestre, 2012; Reddy, Hall & Sulaiman, 2012).

The majority of business models studied were aimed at bottom of the pyramid consumers, but only slightly outnumbered the business models that included the poor as producers. Linna (2012), for example, described how the mobile phone services of Virtual City aimed to improve the agricultural value chain in Kenia, and Hall et al. (2012) studied the effects of increased tourism from Western countries on local Brazilian hotel and shop keepers. Contrary to the examples of Kolk et al. (2012), most of the entrepreneurial activity of the poor that was described was not introduced by the (multinational) company’s bottom of the pyramid initiative; the initiative merely enhanced existing entrepreneurial activity. Examples in which the poor are cast in the role of entrepreneur by a (multinational) company (and usually primarily used as entrepreneurial distribution channels) were still found, but are a minority in the studies published after 2009. Apart from repeating her Avon example, Dolan (2012), for example, discussed the Solar Sister in Uganda, who took on the role of entrepreneurs to distribute their products to the poor via informal networks like the Avon woman have.

About one third of the publications were specific about the outcome of the case(s) studied and the majority of these only reported the outcome for the firm. Karamchandani et al. (2011) publication for example, contained a series of business examples with varying degrees of success, such as Procter & Gamble’s failure to commercially market its water purification powder and Hollard’s funeral insurance, which gained 600,000 customers in five years, and included several explanations as to why the initiatives where successful or not.

Other authors also contrasted successful and unsuccessful bottom of the pyramid cases to demonstrate the importance of the specific aspect of the bottom of the pyramid market they studied (e.g. in Reddy et al.’s (2012) study of innovation vehicles and Wentzel et al.’s (2013) study of sales

(28)

channels). Social outcomes were reported in 7 articles only, and ranged from negative outcomes such as further exclusion of the poorest Nigerians in Sesan et al.’s (2012) Clean-Cook case and increased crime rates in the Brazilian tourism hotspots studied by Hall et al. (2012), to positive outcomes such as the improved standard of living of Indian artisans resulting from the activities of Jaipur Rugs described by Tyagi (2012).

Only 5 articles report an environmental outcome, Hudnut and DeTienne (2010), for example, describe the case of Envirofit, an initiative that reduced greenhouse-gas emissions by motorcycle taxis in the Philippines.

4.2 GF250 characteristics

Out of the 250 companies, 79 reported on bottom of the pyramid initiatives. A total of 147 initiatives were identified, an average of 1.9 initiatives per company. One of these initiatives, SMS for life, was a collaboration between British telecom provider Vodafone and Swiss pharmaceutical Novartis, and was reported by both companies. The majority (n=45) of companies reported only one single BOP initiative, while Novartis reported 7.

(29)

Out of the companies reporting on initiatives at the bottom of the pyramid, the largest number (n=18) originated from the United States (see figure 5). However, as 69 of the Global Fortune 250 companies originate from the United States, this number was relatively low (26%). Only Japanese companies (6 out of 32; 19%) were less active at the bottom of the pyramid, based on their own reports. Chinese companies performed slightly better than American companies (10 out of 33; 30%). The main European countries averaged around 41% (30 out of 74), while the other Asian countries (South Korea, India, Taiwan, Thailand, Malaysia, Indonesia, Singapore and the U.A.E.), although only represented by a small number of companies in the Global Fortune 250, had a relatively high share of companies reporting initiatives at the bottom of the pyramid (11 out of 16; 69%).

From a sustainability reporting point of view, European and Japanese companies were expected to perform better than American companies. The low number of Japanese companies reporting on the bottom of the pyramid was surprising in this respect, however, their reporting on sustainability appeared to be mostly focused on environmental and employee issues and disaster relief, and not on a wider set of social issues such as the bottom of the pyramid, which may explain the difference in their performance on bottom of the pyramid reporting and sustainability reporting in general. The South Korean companies did better than expected, and seem to have caught up with the rest of the world from the low level they were at the time of the study on GF250 sustainability reporting (2005).

Moreover, companies based in several European (Switzerland; 3.7 per company, the United Kingdom; 2.5 per company; Spain, 2.3 per company) and Asian (South Korea; 2.8 per company, India; 2.0 per company) countries reported a relatively high number of initiatives per company.

pyramid; food and beverage production (4 out 7; 57%), chemicals production including pharmaceuticals (7 out of 12; 58%) and telecommunications (8 out of 12; 67%).

(30)

Figure 6: # Companies active at BOP vs. total # of companies in GF250 by industry

In terms of industries, the relative share of companies reporting bottom of the pyramid initiatives for most industries did not deviate considerably from the average (32%; see figure 6). However food- and wholesale stores and construction companies (n=22) did not report any bottom of the pyramid initiatives. In three industries over 50% of companies reported being active at the bottom of the pyramid; food and beverage production (4 out 7; 57%), chemicals production including pharmaceuticals (7 out of 12; 58%) and telecommunications (8 out of 12; 67%).

Moreover, chemical/pharmaceutical (21 by 7; average of 3.0 per company) and telecommunications (22 by 8; average of 2.8 per company) companies reported a relatively high number of initiatives per company.

From a sustainability reporting point of view, companies from the chemical and pharmaceutical industry were expected to be the ones to report most frequently on bottom of the pyramid activities, followed by other smokestack industries. Retail and wholesale companies were expected to be the least frequent reporters. This is consistent with the results on bottom of the pyramid reporting.

(31)

Table 1: Variable means, standard deviations and correlations

Descriptive statistics and correlations for the quantitative variables are reported in table 1. There is a significant (p < .05) positive correlation between return on assets and multinational activity, a highly significant (p < .01) positive correlation between UN Global Compact membership and multinational activity, and a significant negative correlation between return on assets and total assets, however, all of these correlations are relatively weak (.126, .186 and -.147). There are highly significant correlations between UN Global Compact membership and the dependent variables bottom of the pyramid reporting and number of bottom of the pyramid initiatives reported and between the dependent variables themselves, which have no effect on the validity of the regression model.

BOP Initiatives Reported

Variable B SE B Beta (β) Step 1 ROA (%) .005 .005 .073 Total assets ($mm) .000 .000 .059 Step 2 ROA (%) .005 .005 .067 Total assets ($mm) .000 .000 .057 Multinational .068 .085 .051 UNGC Member .222 .058 .239**

Note: N = 250. R2 = .01 for step 1; ΔR2 = .06** for step 2. ** p < .01.

Table 2: Results of Firm Size, Profitability, Multinational Activity and UNGC Membership as predictors for BOP Initiatives Reported

Variable M SD 1 2 3 4 5 6

1. ROA (%) 2.88% 6.32% -

2. Total assets ($mm) 374.22 641.83 -.147* -

3. Multinational (0 = no, 1 = yes) 0.86 0.35 .126* -.056 -

4. BOP Projects Reported (0 = no, 1 = yes) 0.32 0.47 .064 .049 .101 -

5. # BOP projects 0.59 1.16 .094 .002 .085 .748** -

6. UNGC Member (0 = no, 1 = yes) 0.49 0.50 -.005 .018 .186** .249** .223** -

Note: N = 250.

**. Correlation is significant at the 0.01 level (2-tailed).

(32)

# of BOP Initiatives Reported

Variable B SE B Beta (β) Step 1 ROA (%) .018 .012 .096 Total assets ($mm) .000 .000 .016 Step 2 ROA (%) .017 .012 .093 Total assets ($mm) .000 .000 .014 Multinational .114 .212 .034 UNGC Member .502 .146 .217**

Table 3: Results of Firm Size, Profitability, Multinational Activity and UNGC Membership as predictors for the number of BOP Initiatives Reported

Firm size and profitability had no significant effect on either bottom of the pyramid reporting or the number of initiatives reported (p > .05; table 2 and 3). Whether or not a company was a multinational did not have any significant effect on either bottom of the pyramid reporting or the number of initiatives reported. Only United Nations Global Compact membership proved to be a highly significant (p < .01) predictor for reporting on bottom of the pyramid initiatives and the number of initiatives reported.

The adjusted R2 _{of the linear regression models were only .06 and .05 however, leaving a large part}

of the variation in bottom of the pyramid reporting unexplained.

As only 2 cases were collected from the United Nations Global Compact website, and both companies concerned also reported (other) cases on their own website, the use of the United Nations Global Compact as a source therefore cannot (fully) explain the significance of United Nations Global Compact membership as an independent variable.

(33)

4.3 GF250 initiatives

4.3.1 Bottom of the pyramid definition

When analyzing the reported bottom of the pyramid initiatives in detail, several things stood out.

Table 4: Frequency of target group definitions

First of all, only Novartis used the term “bottom of the pyramid” for a target group, and defined this group as the rural poor in India. For other initiatives aimed at people living under similar conditions and at similar income levels outside India, Novartis uses the term “the poor”. Both Argentinian financial Banco Bilbao Vizcaya Argentaria and American pharmaceutical Pfizer use the term “base of the pyramid”, but Pfizer too uses “the poor” and “base of the pyramid” interchangeably. The target group is most often described in terms of being poor; (people living in) “poverty”, the “poor” or the “poorest” (see table 4). The term (people living on a) “low-income” is used to define the target group in 51 of the reported cases.

Out of the initiatives aimed at (people living on a) “low-income”, the majority (55%) took place in developed countries, while the vast majority (86%) of initiatives aimed at target groups described in term of poor/poverty took place in developing countries, which indicates that these terms do not define the same group of people.

Only 12 companies defined their target group in monetary terms (see figure 7). The majority of these income levels would meet Prahalad and Hart’s (2002) definition (< US$ 1,500 annually at PPP); the US$20,000 ceiling defined by American financial Fannie Mae, however, will hardly meet any definition of the bottom of the pyramid.

Target group definition # of initiatives

Poverty 71 Low-income 51 Poor 21 Base of Pyramid 2 Bottom of Pyramid 1 Poorest 1 Total 147

(34)

Figure 7: GF250 target group position in the pyramid (adapted from “The fortune at the Bottom of the Pyramid” by Prahalad and Hart, 2002)

Both British financial HSBC Holdings and South Korean electronics producer LG Electronics refer to the United Nations definition of US$1.25 a day at purchasing power parity, and only China National Offshore Oil defines an income level lower than that (300-600RMB a year at 1995 PPP-levels).

In addition to defining the target group in terms of poverty or income, 48 of the 147 reported cases are aimed at a specific (sub)group: 19 are aimed at either farmers, herdsmen or fishermen, putting the Confucian proverb 'give a man a fish he eats for a day, teach a man to fish he eats for a lifetime’ into practice. Another 14 initiatives were specifically aimed at women, and another 13 at children or youngsters (and 1 at elderly). These target groups are particularly relevant from a development economics point of view, as gender inequality and youth unemployment have been found to negatively affect economic growth.

In 108 of the cases, the specific location of the target group was specified (other cases used geographical terms such as “global”, “developing countries”, etc.).

(35)

Figure 8: Countries hosting GF250 BOP initiatives

The main countries hosting bottom of the pyramid initiatives (see figure 8) were China (n=17) and India (n=14). Considering that over 50% of the people living under the US$2 a day (at PPP) poverty line live in China and India (Sumner, 2012), these results are not surprising (broadening the $2 a day poverty definition to e.g. Prahalad and Hart’s (2002) bottom of the pyramid definition is unlikely to yield a different conclusion).

The third country in which over 10 initiatives took place was the United States (n=12), which is surprising since very little (<0.1% of the total) people below the US$2 a day poverty line live in high-income countries (World Bank definition) such as the United States (Sumner, 2012). These initiatives therefore most likely reached target groups with a higher income level, with the features of Rangan et al.’s (2011) subsistence and low income segments. Although the lowest income groups in developed countries earn more than US$2 a day, they usually struggle to make ends meet from their income from temporary or semi-permanent employment in both the formal and informal market (Wacquant, 2007). The same holds for all other high-income countries hosting bottom of the pyramid initiatives, such as the United Kingdom (n=2) and Australia (n=1).

(36)

Apart from China and India, the countries with the highest number of poor people (15% of world’s total <US$2 a day; Sumner, 2012) are Pakistan, Nigeria and Indonesia. With Nigeria and Indonesia only hosting one initiative each, and Pakistan hosting none, these countries are clearly underserved by the GF250. Latin America on the other hand is overserved, hosting 21 bottom of the pyramid initiatives while only being home to less than 3% of the world’s poor (<US$2 definition; Sumner, 2012).

Out of the 12 initiatives that took place in the United States, 10 were reported by American-based companies, providing a possible explanation for the relatively (vis-à-vis the number of poor) high number of initiatives located in the United States. The same does not hold for the overserved Latin American countries, as none of the Latin American initiatives were reported by Latin American companies. Furthermore, 16 out of the 17 Chinese initiatives were reported by local companies versus only 4 out of the 12 of Indian initiatives, making India the largest net recipient of bottom of the pyramid initiatives.

4.3.2 Business model

In terms of business models, 4 types of models were often reported (see figure 9).

Figure 9: BMs in Organizing Framework (adapted from “Reviewing a decade of research on the “Base/Bottom of the Pyramid” (BOP) concept” Kolk et al., 2012)

Consumers Co-inventors

Adapted BM

Recipients

Producers

Position in value network

M ode o f E ng ag em en t Non-adapted BM Employees A B C D E F G H

(37)

The most common model (group A, n=48) was the model in which the poor were consumers of adapted products or services. French telecom provider France Télécom, for example, offered a text-to-speech and audio mailbox service to illiterate consumers in Africa; Swiss food and beverage producer Nestlé sold Maggi Masala, a single portion of spices with added vitamins and minerals, in India; and American telecom provider Comcast provided free I.T. training with its (low-cost) internet access for low-income American consumers. 7 of the group A cases were either microfinance or microinsurance initiatives aimed at consumers, all of which were reported by banking or insurance companies.

The second most common business model was the model in which the poor were producers within an adapted business model (group B, n=32). Japanese food commodities trading company Itochu focused on farmers in Mozambique as producers in an adapted business model, by not merely buying their supplies but working together to improve growing techniques. Another example is Vodafone, which trained local Qatari women in selling Vodafone products. Microfinance or microinsurance initiatives aimed at producers were also common (n=11), of which 9 were reported by banking or insurance companies.

In the cases were the poor were consumers of a free product or service (group C, n=31), the product or service provided was not related to the company’s general business model in all but one case (filling prescriptions for free by the American pharmacies of CVS Caremark). German car manufacturer Volkswagen, for example, built public water pumps in remote areas in Brazil, and South Korean utilities company Korea Electric Power paid for a 100 patients to have eye surgery in several countries. 11 of the initiatives were reported by Chinese companies and aimed at target groups in China, such as the Uyghur (by metal company Baosteel Group) and Tibetan (by utilities company China Huaneng Group) minorities.

“Recipient” cases in which poor producers received free products or services were not uncommon either (group D, n=13); these cases were a mix of business model and non-business model related.

(38)

American computer manufacturer Hewlett-Packard, for example, provides free e-learning courses for poor people aspiring to be entrepreneurs, while LG Electronics supports Ethiopian farmers in reclaiming farmland. The observation that most recipient models are not business model related is important because giving away business model related products or services for free can be part of a business strategy (Osterwalder & Pigneur, 2010). Hewlett-Packard, for example, may offer free courses on how to use computers with the intent to boost demand for its products, which is an activity different in nature to Korea Electric Power’s eye surgery provision as this activity will not have a direct impact on demand for electricity in Korea.

Only a low number (group E and F, n=5) of non-adapted business models were reported. 3 out of these 5 cases are reported by pharmaceutical companies: Bayer from Germany, Pfizer from the United States and Novartis all report proving (agro-)medicine to the poor via a non-adapted business model.

The poor were rarely cast in the role of employee (group G, n=3). One of the 3 cases was Swedish car manufacturer Volvo, which offered internships to young people vulnerable to HIV in South Africa.

Only one company, American pharmaceutical company Johnson & Johnson, reported actively engaging the poor in the design of the initiative (group H); they co-developed their health training with local Indian women.

In 57 of the 147 cases a non-commercial partner was involved. In 33 cases this partner was a (number of) NGO(s). In all but 3 of these cases the partnership was formed by a company based in a high-income country executing a bottom of the pyramid initiative in a low- or middle-income country (World Bank definition).

Australian mining company BHP Billiton for example, teamed up with international NGO Global Communities to improve living conditions in resource-rich areas in Columbia, hoping to reduce violence and illegal activities in these areas.

(39)

In 19 of these cases, the partner was the local government. For example, German insurance company Allianz reported the need to partner with the Chinese authorities to bring its microinsurance products to the Chinese market. In 5 other cases, a partnership with the Chinese (local) government was reported as well.

In 5 cases, partnerships were formed with both the local government and NGOs. Spanish Telecom

provider Telefónica, for example, teamed up with a number of American and Brazilian (non-)governmental organizations to provide Brazilian fishermen with real time info on weather and

market conditions.

4.3.3 Outcome

In 66 of the 147 cases, some kind of socioeconomic outcome for the target group was reported, however, these outcomes were often (n=26) described in general and hard to quantify terms such as “transforming lives” reported by British mining company Rio Tinto Group or “skill development” reported by Austrian petrol company OMV Group.

In 19 other cases, the outcome was described in numbers of people, households or countries reached, such as the 2100 Cambodian homes supplied with solar energy by South Korean petroleum company GS Caltex, the 27.000 jobs created in the pharmacies of CVS Caremark, and the elimination of the contagious eye disease Trachoma in 3 countries as a result of the efforts of Pfizer. In 9 cases, the outcome was quantified in monetary terms, such as the €190 million increase in agricultural production achieved by the Turkish members of Vodafone´s Farmer´s Club, and the 142% increase in Tibetan household income resulting from improved infrastructure as reported by Chinese petrol company Sinopec Group.

In addition to the socioeconomic outcome, a positive environmental outcome was reported in 17 cases. In 13 of these cases, this positive environmental outcome was a result of a reduction in energy consumption or an increase in the supply of renewable energy such as solar energy. Two cases, reported by American beverage producers Coca-Cola and American postal company FedEx, were

(40)

aimed at promoting recycling on Brazil´s landfills. The other two cases, reported by Itochu and American I.T. company Schlumberger, were aimed at promoting sustainable agriculture in Madagascar and India.

25 cases reported improved health (care) as a goal and/or outcome of the initiative, of which 15 were reported by health care providers or pharmaceutical companies, such as Swiss pharmaceutical company Roche Group, which used a zero-profit pricing scheme for HIV drugs in developing countries. An example from a non-health-care company is the case of France Télécom, which set up a data and communications system between local clinics and larger (more knowledgeable) medical centers in Senegal.

Only South Korean electronics producer Samsung Electronics reported the business outcome of its bottom of the pyramid initiatives: a 184% increase in sales of LED TVs and a 23.5% market share in the African refrigerator market as a result of its product adaptations specifically for the African market.

(41)

5. Discussion and Conclusions

When comparing the bottom of the pyramid initiatives reported in the literature vis-a-vis the bottom of the pyramid initiatives reported by the world’s largest companies themselves, several things stand out.

5.1 Bottom of the pyramid definition

First of all, neither business scholars nor companies seem to agree on the exact definition of the bottom of the pyramid, indicating that one of the main critiques on the bottom of the pyramid (most notably Karnani, 2007) still has not been resolved.

Even though London and Hart (2010) argue for a broader and more flexible set of characteristics to define the bottom of the pyramid, the definition is now too divergent, resulting in business scholars studying different bottoms of the pyramid (Kolk et al., 2012), and companies targeting distinctly different target groups, the difference between the HSBC Holdings’ US$1.25 a day and Fannie Mae’s US$20,000 a year target group being a case in point. This makes a comparison of published articles or self-reported business initiatives, or a comparison between the two, imperfect and this is still one of the main areas for improvement of the bottom of the pyramid concept.

5.2 Company characteristics

The GF250 that reported being active at the bottom of the pyramid, are not just the “usual suspects” studied in the literature. The well-known Western multinationals did report on the same and similar cases to the ones discussed in the literature, such as Unilever’s Shakti project and Nestlé’s Maggi Masala. The non-Western companies active at the bottom of the pyramid in countries other than their country of origin, however, are almost completely ignored in the case studies selected by business scholars, while these initiatives are large in number and scale, such as Itochu’s ProSavana agricultural project in Mozambique and Samsung Electronics’ Triple Protection ACs and TVs sold in Africa. Local non-Western initiatives are selected as case studies, but focus mostly on Indian (and to

Beyond shampoo, salt and SIM-cards : an analysis of BOP initiatives reported by the Global Fortune 250.