Introduction
The response to the HIV/AIDS epidemic has to date been dominated by biomedical and behavioural interventions (Campbell and Williams 1999), with moderate but disappointing progress in relation to sexual behaviour change (Padian et al. 2010). Although it has long been acknowledged that context and social structures have influenced the dynamics of the epidemic (Hahn 1991; Sweat and Denison 1995), it is only more recently that these insights have been fully incorporated into the global policy agenda (UNAIDS 2010), in part as a result of a growing body of literature that has emphasised the structural or social drivers of HIV (Sumartojo 2000; Auerbach et al. 2010; Auerbach, Parkhurst, and Cáceres 2011). This is a much-needed development which transcends early transmission models which are generally individualistic in nature (King 1999). Transactional sex has been identified as one of the most important structural drivers of HIV risk (Jewkes et al. 2012). However, transactional sex has often been defined differently and most of these definitions have not been comprehensive enough to capture the dynamics of the practice in a sub-Saharan African setting. Broadly speaking, the term ‘transactional sex’ refers to sexual interactions in which something is exchanged or transferred, though on a more informal basis than, and conceptually distinct from, commercial sex work (Hunter 2002; Dunkle et al. 2004; Jewkes et al. 2012). A more formalised definition suggests that transactional sex should be defined as ‘a sexual relationship or act(s), outside of marriage or sex work, motivated primarily by the expectation of material gain, where love and trust are also sometimes present (involved/concerned/at play)’ (Stoebenau et al. forthcoming).
Whilst having sex is not in itself risky, transactional sex is frequently associated with intimate partner violence, situations in which women are often unable to negotiate condom use, and intergenerational sex which further exacerbates power differentials over the terms and timing of sexual interactions (Dunkle et al. 2004; Jewkes et al. 2012). Statistical work confirms that female participation in transactional sex enhances the likelihood of HIV infection (Dunkle et al. 2004; Jewkes et al. 2012). Although there are some dissenting voices (Stillwaggon and Sawers 2012), the statistical evidence, in light of insights from qualitative work, strongly suggests that this behaviour may be risky and that it needs addressing in current and future prevention efforts. However, despite general agreement that the practice of transactional sex is rooted in unequal and gendered power relations, it is not always clear exactly how transactional sex is conceptualised in a structural way, and how sexual interactions that involve exchange reflect the dynamics of the social system and broader historical socio-economic processes.
Although there are a range of excellent analyses of the social complexities around the practice of transactional sex (Poulin 2007; Swidler and Watkins 2007), mainstream economic attempts to address this issue have to date largely failed to provide useful insights, reflecting a general lack of a substantive contribution by the economics profession to the epidemic (Johnston 2013). In most cases, economic analyses are rooted in the standard microeconomic framework in which individuals make rational decisions over their sexual behaviour, an approach that fails to respond to theoretical and empirical work originating from other disciplines regarding the role of structural drivers, or engages with them to the limited extent permitted by the underlying technical apparatus (Milonakis and Fine 2008). For example, early attempts to model sexual behaviour explained the continued risky sexual behaviour of poor sex workers in light of the epidemic through the prism of the opportunity cost, as sex workers weighed up the perceived probability of infection against the material gains of a given sexual encounter (Philipson and Posner 1995). The opportunity cost is central to later more technical work, though with application to the wider population rather than specific high-risk groups, in which individuals make sexual decisions based on their current and expected lifetime utility, with the immediate lost utility of not having sex related to anticipated future utility (Oster 2012). Transactional sex has also been framed in a similar way (Luke 2006; Robinson and Yeh 2011; De Walque, Dow, and Gong 2013), with attention to the existence of a premium for unsafe sex, and how female decision-making with regards to sexual behaviour responds to external shocks. Whilst these analyses provide some useful insights, overall they fail to address central concerns related to transactional sex, especially unequal and gendered power relations, reflected in the broader literature (Bene and Merten 2008; Jewkes et al. 2012; MacPherson et al. 2012), and which present a strong challenge to models framed in the language of free choice.
Despite these limitations, the mainstream economic framework underpins current interventions such as conditional cash transfers (Baird et al. 2012) and microfinance (Pronyk et al. 2005), which are in part designed to reduce the need for women and girls to engage in transactional sex, and which are currently gaining prominence as ‘structural’ interventions (see O'Laughlin in this issue). This reductionism of scope and level of analysis, with individualistic economic theory (re)incorporating the social in a limited way (Fine and Milonakis 2009), highlights a pressing need for alternative economic approaches that can address broader concerns around power and gender in relation to transactional sex.
This article will reflect on the economics of transactional sex, using evidence from a research project conducted in Tanzania to develop alternative economic conceptualisations and explanations of this practice. This will complement and add to the growing body of work that roots this practice in the workings of the economic and social system. Our research illustrates the range and complexity of transactional sexual interactions, and illustrates how an alternative economic approach can shed light on the structural dynamics of transactional sex. This enables a critical reflection on current economic approaches to HIV prevention targeted at transactional sex, and implications for a different policy agenda that addresses women's empowerment more generally. The rest of this article is structured as follows. In the next section, we critically assess mainstream approaches to transactional sex, before introducing the study site and fieldwork methods and then presenting evidence from our fieldwork. This enables the formulation of an alternative economic understanding in the following section, and a discussion of the implications for prevention policy that precedes some concluding remarks.
The economics of transactional sex
Mainstream economic approaches to the study of transactional sex, such as Luke (2006), Robinson and Yeh (2011), and De Walque, Dow, and Gong (2013), are articulated in the language of rational choice, in which sex and either money or non-monetary gifts or goods are traded in a market setting (Luke 2006; Robinson and Yeh 2011; De Walque, Dow, and Gong 2013). Typically, survey data is used to statistically test hypotheses that reflect underlying behavioural models. The study of sexual practices that involve some form of exchange using this framework inevitably focuses on the monetary value of whatever has been exchanged for sex, which enables the construction of a market for sex/safe sex, in which the value of the transaction is related to the likelihood that a woman will engage in transactional sex, and/or the likelihood that a condom will or will not be used.
Luke (2006) collected data on informal sexual relationships in urban Kenya, and sought to test whether a market for safe sex existed in which women with given preferences and wealth endowments choose from a range of different possible partnerships ‘where each partnership is characterized by a level of transfers and a probability of condom use’ (322). The underlying model then tests whether the probability of condom use is related to the size of the transfer, with the hypothesis that condom use will be negatively related to the size of the transfer confirmed as statistically significant. This result is interpreted as evidence that women are active agents in the process of negotiating condom use within transactional settings, and confirmation that informal sexual relationships that involve exchange can be conceptualised as a functioning market for unprotected sex. A further implication of this work is that there is a ‘premium’ for having sex without a condom, a notion taken up by Robinson and Yeh (2011), who seek to understand whether the existence of a premium for unsafe sex enables women to increase their supply of unsafe sex in times of need. The existence of a risk premium is tested by examining whether the ‘price’ of a sexual transaction is related to risky sexual activities, which is confirmed as statistically significant, and thus ‘it may be rational for women to choose to engage in unprotected sex to capture the risk premium’ (Ibid., 50). Having established a clear motivation for women to engage in transactional sex, they then go on to test whether unprotected sex is related to household shocks through two fixed-effects equations that also test whether households maintain consumption after a shock by reducing other expenditures. The statistical results confirm that these women increase their supply of unsafe sex as a reaction to relatively small household shocks, again confirming that sexual interactions involving exchange are underpinned by the logic of the market. These findings echo those of similar modelling exercises that focus on condom use by commercial sex workers (Gertler, Shah, and Bertozzi 2005).
Finally, De Walque, Dow, and Gong (2013) using panel data from rural Tanzania on women whose primary means of survival is agriculture, find that following a negative household shock, both married and unmarried women have more unprotected sex, and are also more likely to have a sexually transmitted infection. In this case, transactional sex is framed as a way of coping with adverse shocks, and this study highlights that this does not just apply to single, divorced or separated women, but is a practice engaged in by the wider population. This study also links unprotected sex to the risk of contracting a sexually transmitted infection, and thus a direct impact on the potential for HIV transmission. As with the other approaches discussed here, women are viewed to increase or decrease the amount of unprotected sex that they have depending on their material needs, and to some degree this captures the notion of ‘survival sex’, in which poverty and desperation drive women to engage in sex of a transactional nature (Wojcicki 2002).
There are of course some differences between these approaches. In particular, whilst Luke (2006) uses a broader definition of transactional sex that seeks to assess informal relationships and exchange, and De Walque, Dow, and Gong (2013) focus on the general female rural population, Robinson and Yeh (2011) sample women that are ‘single, widowed, divorced, or separated woman, aged 18 or older, who had multiple concurrent sex partners’ (40), which is not only a very specific target group, but also risks conflating commercial sex work and transactional sex, a blurring of sexual interactions that other authors have been at pains to conceptualise as distinct (Hunter 2002; Leclerc-Madlala 2003; Jewkes et al. 2012).
These economic approaches provide a number of useful insights into the mechanics of transactional sex which echo themes in the broader anthropological and public health literature. Women are not characterised as passive victims, but are ascribed a degree of agency, thus challenging unhelpful negative stereotypes of African women (Hunter 2002; Leclerc-Madlala 2003). The pernicious role of poverty is also brought to the fore, and in particular the vulnerability of fragile households to external shocks. The authors also suggest a number of avenues for intervention that are echoed in the broader literature and that implicitly acknowledge the constrained socio-economic context within which many women live, such as improving access to health care, and especially health care for dependents (Robinson and Yeh 2011), the provision of alternative economic opportunities for women so that they do not have to resort to exchanging sex for survival (Luke 2006), and the need to financially empower poor and vulnerable women (De Walque, Dow, and Gong 2013). These are welcome additions to strategies that go beyond the standard package of biomedical and behavioural interventions (Campbell and Williams 1999).
However, there are a number of limitations associated with these economic approaches that are related to the economic framework employed, which emphasise the degree to which these analyses are at odds with the broader literature. First, whilst there is some acknowledgment of the challenges that women face, it is unclear whether the notion of agency, conceptualised as the ability for women to rationally and freely choose whether or not to engage in transactional sex, reflects the influence of unequal economic and gender relations that frame these decisions. Indeed, the use of the loaded term ‘choice’ in this context is certainly questionable (Johnston 2011), with the options that poor(er) women face within a constrained socio-economic environment more akin to the frying pan or the fire rather than alternative outcomes that can be regarded in any sense as optimal (Christensen 1998). This narrow formulation of agency ignores issues such as power, force and coercion. A cursory word search of the three papers finds that the term ‘coercion’ does not appear in any of them; the word ‘power' does twice in Robinson and Yeh (2011) and De Walque, Dow, and Gong (2013), though in both papers on one occasion this refers to the power of the statistical test, and the word ‘coerce' appears once in a footnote in Luke (2006). Further, the word ‘force’ only appears in Robinson and Yeh, though not directly in relation to sexual behaviour. These are not linguistic omissions, and represent a systematic failure to address concerns that fall outside of the market framework which explicitly does not engage with questions of unequal power, instead assuming free market entry and, importantly, exit.
This is in stark contrast to the broader literature on transactional sex, in which unequal gender relations, power and coercion are central to the analysis. MacPherson et al. (2012) examined transactional sexual interactions between fishermen and female fish traders in Malawi, finding that fishermen used their economic position to extract sex from traders who were desperate to ensure that they could access fish, particularly in the lean seasons, and in general ‘exploit women's economic need by pressuring them into having sex with them’ (7). Stoebenau et al. (2011), whilst acknowledging that women actively use their sexuality to extract material goods from men, also note the importance of power, not only in relation to the highly unequal gender relations which provide the backdrop for these sexual interactions, but also ‘at the point of the sexual encounter, where men typically determine the terms and, in some cases, do so with violence' (12). These insights emphasise that power and coercion in transactional sex influences both whether women participate, and also the nature of the sexual interaction, two different elements that are conflated in the standard economic framework.
A related concern is the focus on the quantity of money or equivalent value of goods exchanged as the primary motivating factor for women to engage in (unprotected) transactional sex. This is not unsurprising given that these are primarily economic analyses. However, a range of other important factors such as local norms, customs and obligations which shape transactional interactions and emphasise that the value of the exchange can often be a secondary consideration, remain unaccounted for (Hunter 2002; Poulin 2007; Swidler and Watkins 2007). This focus on the exchange element leads to a narrow range of economic interventions, such as conditional cash transfers for schoolgirls which seek to reduce the need for them to engage in transactional sex by providing the income that is seen to be the initial motivator (Baird et al. 2009; Fiszbein et al. 2009), or microfinance, which aims to reduce female reliance on transactional sex as a source of vital income (Pronyk et al. 2005). These approaches fail to incorporate a more nuanced understanding of transactional sex, ignore factors such as norms and conventions which may in fact be a more pertinent point of intervention, and in some cases may make things worse (see MacPherson et al. 2015).
A final limitation is that these approaches are framed implicitly within a poverty narrative, particularly in relation to the response by poor women to adverse household shocks. Whilst this is no doubt an important driver of some transactional interactions, the broader literature emphasises that transactional sex is not just engaged in for basic survival goods, but also for consumption goods or status goods (Hunter 2002; Leclerc-Madlala 2003), and not just by women that are poor (Chatterji et al. 2004). Indeed, it is noted that processes of globalisation, consumerism and the associated expansion of needs that are related to capitalist development may in fact enhance the pressures for women to engage in transactional sex (Hunter 2002; Leclerc-Madlala 2003; Dunkle et al. 2004; Stoebenau et al. 2011), rather than development bringing about a reduction in motivations for transactional sex by fulfilling basic material needs. This highlights the dynamic nature of sexual relations and analytical categories, and also how structural economic processes play an important role in shaping individual behaviours.
This discussion emphasises the limitations and omissions of mainstream approaches to transactional sex, and also how divorced these approaches are from the themes highlighted by the broader public health literature. In particular, transactional sex is viewed in a stylised manner, with little incorporation of issues such as gendered and economic power and coercion, and how transactional sexual interactions are framed by structural dynamics. The economic element is also extremely narrow, reduced to the element of exchange. This ignores the many different forms of transactional sex in practice, and thus presents an incomplete analysis. Despite the dominance of the rational choice framework in the economics profession, there are, however, other approaches which address notions of inequality and power, and which help shed light on an alternative economics of transactional sex.
Fieldwork methods
The evidence presented below is taken from data gathered in a qualitative research project conducted in Mwanza region, northwestern Tanzania, which investigated the relationship between temporary economic population mobility and HIV risk. The main component of the fieldwork was comprised of three interlinked phases. In the first phase, four focus groups were conducted to select the mobile groups to be studied in the rest of the research, using a participatory ranking process that aimed to identify the most important forms of mobility engaged in by the local community. The participants in this stage were men and women from both rural and urban areas within the study site, with focus groups conducted with men and women separately to try and ensure that women's experiences were captured. The mobile groups selected in this stage were mobile farmers, maize traders and dagaa 1 sellers. The second phase involved process-mapping exercises with a sample of each mobile group again in a focus group setting, with participants exploring issues of mobility and sexual behaviour, though in relation to the ‘general’ process, including discussions around the systemic and individual factors that shaped patterns of movement and specific destinations, and the nature of sexual interactions engaged in by mobile individuals. In the third phase, a series of in-depth interviews were conducted with a sub-sample of each mobile group to understand participants' own experiences of engaging in specific forms of mobility, including questions relating to their own sexual behaviour. Owing to emerging themes relating to the issue of transactional sex, two additional focus groups were conducted towards the end of the fieldwork to explore local sexual norms in more detail. All of these activities were conducted in the local languages, Swahili and Sukuma, by local research assistants who were the same sex as the participants.2 A thorough debrief was conducted after each research activity, and later complemented with translated transcripts for full analysis and triangulation. Other concurrent activities conducted to understand more about the local socio-economic setting, and to contextualise the qualitative component, included interviews with key informants such as local village and government officials, informal discussions with the local research community, the author's own observations and further secondary research.
Whilst transactional sex was not the initial focus of the study, it was a theme that emerged during the research process, and in particular when mobile individuals were talking about their sexual interactions while they were away. Although mobility may have enhanced the opportunities for doing so, it was also clear that transactional sex was not just engaged in by those who were mobile, and indeed this led to a reconsideration of the role of mobility as the project progressed, with the findings suggesting that sexual interactions were strongly influenced by gender relations and local sexual norms rather than mobility. This shed light on how risk was shaped for both mobile and non-mobile populations, and suggested one explanation as to why previous statistical work found few differences in either sexual behaviour or HIV prevalence/incidence between mobile and non-mobile groups (Deane, Parkhurst, and Johnston 2010), hence the focus on transactional sex in this paper. The themes that we present here have important implications for the way that transactional sex is conceptualised in economic terms, discussed in subsequent sections.
Transactional sex in northern Tanzania
Transactional sex as a norm
One key theme reported in both focus groups and interviews was that transactional sex, as we have defined it above, is an established and accepted social norm, with broad agreement within the community that there is an expectation of some form of exchange or transfer when extra- or non-marital sex takes place. For example, male maize traders reported that, in relation to opportunities to have sex whilst they were away, ‘when you have … little amount of money … you can look for a friend … but if you have run out completely … it is impossible,’ emphasising that some form of transfer was expected. In later focus groups conducted with male and female adult participants to explore the issue of transactional sex, participants agreed that this was their expectation:
You can surely tell him and he accepts everything, but in his heart, he says if I say I do not have [money], I cannot go with this woman [to have sex]. (Female focus group participant)
R:Eeh, it's primary. Even if she's given something else money is the primary thing. (Male focus group participant)
This expectation reflects similar findings in research by Wamoyi et al. (2010, 2011) with young people and their parents, conducted in the same study site, in which parents noted that a girl's private parts are like a ‘shop’ (‘hayo ni maduka’). This expresses the view that nothing is obtained from a shop for free, and hence when applied to sex, women have to be given something in return for satisfying male sexual desires. In fact, fathers argued that sex should never be free as it would make it difficult for men to get women, as no woman would agree to have sex for nothing in return, and parents and grandparents all expressed the view that if a man has sex with their daughter or granddaughter without giving her anything, they will ‘have made a fool of her’. Expectations around exchange and casual and informal sexual relationships are also reported in other studies on transactional sex conducted in Tanzania (Maganja et al. 2007).
Operationalising transactional norms
Transactional sexual interactions are engaged in by a wide range of actors in an array of divergent socio-economic contexts, for multiple reasons, and hence avoid neat categorisation. One common theme is that men can utilise transactional norms to stake a future claim on sex by giving gifts upfront, in a sense a down payment which is then followed up on and used to coerce sex at a later time or date. This may involve buying someone a soda, or giving them a few extra tomatoes or fish at the market:
You may be bought a soda, someone may pass selling mangoes you may be bought and you eat, when it gets to the evening, the one who was buying you those things starts to follow you [ask for sex] because of what he gave you. (Female seller of dagaa [small fish, see note 1])
This phenomenon has previously been reported in relation to younger women still at school, with gifts such as sugar cane, vitumbua 3 or sweets often received from older men as a means of pressuring them into sex (Wamoyi et al. 2010).
Conversely, men report that they feel targeted by women, who seduce them primarily to gain access to their money. In one example, a male maize trader noted that he is approached by women on buying trips at times when he has available cash to hand, whilst one farmer reported feeling targeted owing to the fact that if his female employee convinced him to have sex with her, he would have to give her something, such as a wage increase or one-off payment:
Now if you keep women there must be temptations. Those women will want sex from you so that you can sometime increase the amount of money, you may have agreed, she seduces you so that you may increase the money through sex. (Male farmer)
Whilst these may in part be narratives used by men that seek to absolve themselves of any responsibility for the extra-marital sex that they engage in, this captures the notion that women are often active agents, albeit in a constrained socio-economic gendered context, rather than merely passive victims, in transactional interactions (Hunter 2002). Further, the manner in which transactional interactions are characterised will also depend on who is giving the account.
Sex and local value chains
Transactional interactions were also reported in relation to participation in local value chains, though again there are different accounts of how this takes place. From a male perspective, male fish traders report giving female customers fish in advance on the basis that they will be paid another time, creating the space for women to attempt to pay back the debt through sex.
Now you may find that someone comes for a loan. You may loan him/her thinking that he/she is a customer but you may find that he/she has some motives. Now in paying he/she may start giving you excuses, now that needs a wise mind, this person is like this and this is his/her motive. (Male dagaa seller)
However, men can also use this situation to extract sex by allowing women to come back and pay later on, and once the goods are loaned and the debt created, they have something to negotiate with:
A man is … is … you ask for a loan, say lend me dagaa, as soon as he lends you, you are tempted. The other day when you send money to him he says just send that money. You meet again later, eeh . . there is no need to pay me back the money, we should just make love. (Female dagaa seller)
In a similar vein, transactional sex is related to a local informal credit practice known as mali kauli, in which traders provide goods upfront to street sellers who then pay them back from the day's proceeds. This enables the distribution of goods across local value networks without the need for daily cash transactions, and also the space for a degree of negotiation and renegotiation of prices as market conditions change, thus protecting traders and sellers from fluctuations in the market in conditions of mutually beneficial ongoing economic relationships (Ogawa 2006). This is another situation in which sex can ultimately be exchanged in lieu of the debt repayment, but within the structure of local trading practices and credit arrangements. This was reported by local maize traders, tomato intermediary traders and dagaa sellers, and is a practice that has parallels with transactional sex related to the workings of the fish value chain observed across sub-Saharan Africa (Gordon 2005; Bene and Merten 2008; MacPherson et al. 2012):
Now that can find you at times when you are sexually aroused. That means that it tempts you because one can buy maize three or four times, but the fifth she decides to seek for a loan, and tells you many things, now you will find out her intentions for coming, so it mean if you also have directed your thoughts there, that's where the business ends … (Male maize trader)
In this situation it is unclear whether sex is engaged in by female sellers because they simply have not managed to sell enough that day to repay the debt, or whether this is a speculative attempt to increase their profits. Further, as with other forms of transactional sex discussed above, the degree of agency and coercion will vary, with some women forced to pay the debt back through sex, whereas some will view these economic relationships as opportunities to supplement their capital. Importantly, it is the gendered nature of these value chains, in which men predominantly sell to women, which provide the circumstances in which gendered and economic power can be expressed. However, it must also be made clear that transactional sex is not always the outcome, as men may not be interested in having sex and will demand repayment through other means such as taking possessions or deferring to a later date.
There are, however, alternative views on the role of sex and exchange and local value chains which reflect unequal gender relations in an entirely different way. One female participant noted that borrowing goods is a way for women to instigate relationships with men in a context in which they are not overtly supposed to do, which emphasises the multiple views and understandings of transactional sex held by those engaging in them:
R:A man hasn't dared to ask for things of love, but she has desires for him, now when she desires him, she will insist on her point that I will go and ask for a loan of something, but today I've got money but I don't want to pay, what I want is …
R:Love. When he comes, ‘I want my money,’ I will be wandering this and that way while I've got the money, what I am targeting there is sex. But if his blood and mine are attracted, he will just ask that I need you, so let's put aside issues related to that money. That's when love/sex starts. Issues of loan are also put aside. (Female focus group participant)
A final and related category is transactional sex engaged in by female businesswomen of different economic status to increase the capital available to them:
Perhaps when you see that the capital is small you decide maybe I should do this, maybe if you have a big capital you cannot ask for a loan from anyone. (Female dagaa seller)
In particular, this quote emphasises that it is not just poor women with little capital who engage in sex for exchange, but also women who are seeking to expand their business. This is a theme that is reflected in other recent work (Hunter 2002; Leclerc-Madlala 2003), and is an extremely important issue if transactional sex is to be better understood, and if prevention efforts are to move away from the poverty narrative that mainstream economic approaches continue to perpetuate.
As noted above, transactional sex has been related to changes in the social system in which pressures for luxury goods and cultures of consumerism, in concert with differential abilities of men and women to fulfil these needs independently, have contributed to the growth of transactional sexual practices. This was again reflected in this study, with focus group participants noting the pressures that women are under in terms of accessing luxury items:
Previously they were just living, today a person may leave here and go somewhere to do her things, meaning living for what … for men, and those men do not have sex with her without bribing, how will she live, how will she eat, what will she wear without being bribed, what makes us being bribed is luxury, thinking of bribery, maybe they were brought by whites. (Female focus group participant)
This further emphasises the need to understand transactional sex as a practice that has specific socio-historical underpinnings, and that, rather than reflecting transactions in an ahistorical ‘market’ for sex, this practice is not inevitable and reflects changes in the social system, and specifically the penetration of capitalism and the monetisation of economic life.
The economics of transactional sex revisited
The evidence presented here illustrates the array of concrete forms that transactional sexual interactions take, which involve multiple and overlapping explanations, varying degrees of gendered and economic power, a range of different motivations, and are engaged in by women from different socio-economic groups. Whilst these findings corroborate previous work on transactional sex, here we reflect specifically on the mainstream economic approaches reviewed above that are primarily organised around the value of the goods or money that is exchanged.
The conflicting reports that men and women give when talking about transactional sex emphasise that the degree of coercion involved will vary, and that women will enter into these transactions with different degrees of awareness and control over what is expected of them. In some cases women report that men give these gifts without any mention of sex, whereas in others, it is unclear to what extent women have the power or economic security to refuse these gifts. This does not preclude the notion that women can accept these gifts in the knowledge that by doing so there is an expectation of them agreeing to have sex later on, or that they are actively engaging in transactional interactions for material gain. As with other accounts that suggest the exchange is often of secondary importance, we also find evidence to suggest that those engaging in transactional sex do so with contrasting motives and aims, with women using informal credit practices as a way of instigating relationships. These complex and overlapping motivations are not accounted for in mainstream, individualistic approaches, and are at odds with mainstream economic views of the market for informal sex in which both parties enter into the exchange with a common understanding of the specific nature of the transaction. Additionally, the amount transferred, and the frequency, vary across interactions, with exchanges or transfers not required in every instance in ongoing relationships, and nor do ‘exchanges’ of sex and money/goods always take place at the same time. This presents a challenge for economic analyses that focus on the amount transferred and assume that the exchange is always related to sexual interactions in the same way.
One central issue is the role of social norms, a current focus within the structural drivers literature. In relation to transactional sex, which we have established as a social norm above, it is unclear to what extent economic approaches framed in the language of rational choice are able to incorporate the role of social norms in shaping sexual behaviour. Indeed, one central critique of mainstream economics is that behaviour is often motivated by a range of factors, including not just social norms, but tradition, custom and habit (Davis 2003). Whilst the mainstream economic approaches reviewed above emphasise the role of agency, this agency can be reconceptualised as a form of agency in which men and women creatively utilise, apply and reapply this norm within a specific and changing socio-economic context. As our research suggests, men and women both actively attempt to use this norm to either extract sex (in the case of men) or for material gain (in the case of women), with varying degrees of success and intent. This supports the notion that women are not just passive victims in this process but may utilise their sexuality for material gain, although this must be understood in relation to dominant norms and socio-economic structures.
However, the acknowledged role of social norms raises the question as to whether social norms around transactional sex have any economic content, or whether they should be understood as purely ‘cultural’ or ‘social’. Our research suggests that the development of transactional sexual norms has been influenced by an increasing commodification of social life, with transactional sex linked to the demand for luxury goods that comes with the development of capitalist relations. Lugalla et al. (1999) argue that transactional sexual practices have replaced older forms of reciprocal exchange associated with goodwill, in the light of intensifying poverty (Ibid.). Transactional sex also reflects the erosion of patriarchal relations to some extent (Wamoyi et al. 2010), as men who want to have sex with women in non-marital situations are now in a position in which they are required to provide something. Other explanations of transactional sex emphasise the historical institutional roots of these practices, such as lubambo,4 which governs the legitimacy of extra-marital relationships in Zambia and which frames fish-for-sex transactions (Merten and Haller 2007). These narratives suggest that whilst the giving of gifts in exchange for sex is a more recent phenomenon, in part a response to increasing economic liberalisation and change, with transactional sex informed by or replacing local institutions or practices, processes related to development may be enhancing risk. These dynamics are not reflected in mainstream economic approaches, which remain ahistorical, with little concern other than how individuals make optimising decisions. Luke (2006) discusses whether the market for safe sex is a recent phenomenon, and whilst acknowledging that ‘the practice of trading money for sex may be a historical occurrence in Kisumu’ (344), concludes that this is in fact a response to the HIV/AIDS epidemic, presumably because the epidemic has significantly altered the potential costs of having sex, opening the space for negotiation and optimising agents' responses to different ‘prices’. However, this ignores how, as noted above, transactional sex is framed by economic and social processes.
Transactional interactions are also related to participation in local value chains, and the highly gendered nature of these value chains, in which gendered interfaces exist where predominantly men sell to women, creating the space in which gendered and economic power can be expressed. Transactional sex can thus occur in a range of contexts and for different reasons. Men can attempt to extract sex through the extension of credit, whereas women can also use informal credit arrangements as a space in which to repay a debt, either because they have not made enough profit, or to expand their capital. This makes clear that transactions that appear the same can in fact be undertaken for entirely different reasons. However, for some, risk will be experienced owing to participation in value chains in which income and profits are often small, variable and subject to daily fluctuations. In comparison to rational choice economic approaches that frame transactional sex as a response to household shocks, we find that risk can be encountered in the daily undertaking of livelihood activities central to processes of household production and reproduction, owing to participation in fragile economic activities.
Following this, transactional sex, as is alluded to in Luke (2006), has a further consequence in that it structures who has sex with whom, particularly as men are required to have access to the necessary resources for the exchange. Our research finds examples of older men having sex with younger women, employer with employee, trader with street seller, creditor with borrower, and businessmen with hotel worker, to list but a few. This has implications for HIV transmission, particularly as this shapes patterns of intergenerational sex, and thus enhanced risk for women. These partnerships illustrate that transactional relationships reflect power relationships that are rooted in prevailing social relations and the intersection of a range of forms of unequal power. This inequality of power is also expressed in the rare(r) occasions in which women are the ones who are gaining sex through giving money to men, something that was touched upon by participants in our study site and noted in other research projects (Dunkle et al. 2007), adding further nuance, suggesting that in some cases economic power is more important than gendered power. These explanations that address power relations, including the influence of gendered and economic coercion, are in stark contrast to the rational choice models that underpin mainstream economic approaches (Christensen 1998).
This brief discussion enables a reflection on different economic approaches to understanding transactional sex. A political economy approach views the economic content as rooted in the role of women in production and reproduction, and locates transactional sex within the workings of the economic system. The forms of transactional sex reported in our study site also reflect other important economic processes, such as increasing consumerism and peer pressure for status among young women in school, or informal credit arrangements in local value chains that are key to enabling the distribution of goods along complex chains with limited cash transactions. This re-emphasises that an economic approach to transactional sex must look beyond the transaction and the value of the exchange for a more comprehensive economic and structural analysis.
Implications for HIV prevention and public health
Our analysis has a number of implications for future HIV prevention efforts and public health. First, following from the notion that the value of the exchange is not always the primary focus of many transactional interactions, and that transactional sex must be located in local sexual norms around sex and exchange, themselves rooted in broader socio-economic relations, the transaction should not be the primary site of intervention. Programmes such as conditional cash transfer programmes and microfinance programmes that aim to decrease women's reliance on transactional sex by giving them either additional income or the opportunity to earn more income are unlikely to succeed, as they are primarily aimed at what we consider to be the wrong target, and based on a limited individualistic analysis. Some small public health enhancements may be achieved in this way, but the broader social relations remain unaddressed. However, if women are given an expanded access to microfinance and income-generating opportunities, the result may be that rather than rebalancing power, their value to men as productive assets increases, further entrenching current gender relations. The expansion of female income generation may also enhance, rather than fulfil, female consumption possibilities, with increased consumption leading to the creation of new wants and needs, with motivations for transactional sex maintained but in a different form. Further, interventions may need to go beyond reducing individual vulnerability and poverty through moderate expansions of income to trying to ensure that the gains from the sustained rapid economic growth that many sub-Saharan African countries are currently experiencing (IMF 2013) are not exclusively captured by men. A related issue is that, as we have argued above, transactional sex is a social norm, rooted in broader socio-economic processes, and with historical roots. This highlights the limitations of behaviour change programmes, and the challenges encountered by incentive-based approaches that attempt to address institutionalised historical practices.
A second key and related theme of intervention, and one that microfinance and cash transfer programmes aim to address, albeit in a limited way, is women's empowerment. It is widely recognised that female empowerment is an essential component of HIV prevention programmes. Reflecting the suggestions of the economic analyses reviewed above, the broader literature on transactional sex puts forward a similar policy agenda, for example the need to improve education, keep young women in school longer, address parental guidance and enhance income-generating opportunities (Chatterji et al. 2004; Bene and Merten 2008; Jewkes et al. 2012). To some extent, different approaches have yielded similar policy recommendations,. This is an issue that has been grappled with across the developmental arena, embodied in the third Millennium Development Goal (Kabeer 2005). One (but not the only) critical approach to how empowerment is conceptualised, and which offers the potential for a more radical avenue of intervention, emphasises that ‘gender inequalities are multi-dimensional and cannot be reduced to some single and universally agreed set of priorities’ (Ibid., 23). In this sense, empowerment is not something that can be achieved through a set of isolated interventions, and requires a deeper degree of social change ‘in which policy changes are implemented in ways that allow women themselves to participate, to monitor, and to hold policy makers, corporations, and other relevant actors accountable for their actions’ (Ibid., 23). For Kabeer, the key issue for international donors is whether they are prepared to fund grassroots women's organisations to mobilise women in the fight for greater equality, a form of collective, not individualistic, action. This is applicable to HIV/AIDS-related interventions, with the key to greater autonomy over sexual interactions, and particularly transactional sex, rooted in women's empowerment through collective action.
Reflecting on the structural drivers of HIV, an application of this approach emphasises that better health outcomes for women are a political issue, and involve a collective struggle against entrenched male power. However, mainstream economic approaches consistently fail to engage with this political element, as behaviour is conceptualised at the level of the individual, with little space for incorporating the need for collective action. There are also concerns regarding how these forms of struggle fit into standard public health and epidemiological preventative frameworks, or with donor priorities for measurable outcomes, in which the number of condoms distributed, or the number of people given access to ARVs (antiretrovirals) can be quantified (Hunsmann 2010). The pressure for immediate results also presents a challenge to funding grassroots women's groups, as the degree of social change required will be an inherently uneven process. Further, it is even unclear to what extent grassroots movements can be beholden to external donors, and whether by doing so the overall aims of grassroots movements are compromised (Beckmann and Bujra 2010). This presents a significant challenge in relation to prevention funding that requires more research and attention.
However, the current consensus that unequal gender relations are one of the underlying drivers of HIV suggests that HIV prevention will necessarily involve the transformation of these relations. The debate then becomes whether this is possible through piecemeal interventions that seek to empower women through small transfers of money in the hope that this will give them greater economic autonomy, or other forms of intervention that are more political and collective in nature (Kabeer 2003; Sweetman 2013). The incentives and market-based approach to social engineering is not only misfocused, but also has no historical precedent. It fails to account for the ways in which women in other settings and time periods have struggled for and taken power, and which have frequently involved women as a social group demanding a greater degree of equality. HIV prevention efforts that seek to empower women can, then, learn lessons from the ways in which women have historically struggled for greater freedom. This locates HIV prevention within broader developmental processes of change, but processes of change that are by no means certain, and that involve challenging entrenched male power and dominance. It will also require a significant reorientation of current prevention efforts and the overcoming of institutional resistance to alternative approaches (Hunsmann 2010).
Conclusion
This article has argued that mainstream economic approaches to transactional sex capture only a limited and stylised view of this practice. In particular, we conceptualise transactional sex as a practice that is underpinned by social norms rooted in the historical development of gender relations. Further, the evidence we present illustrates how sexual norms around transactional sex are utilised by a range of actors across different contexts, and also that transactional norms structure who has sex with whom, and thus are located within the context of broader social relations. This alternative, yet admittedly incomplete, economic approach has illustrated the need specifically for improved understandings of social norms that include an analysis of their economic content. This entails understanding economics in a broader sense, rather than the narrow optimisation framework and focus on scarcity and trade-offs that have come to dominate the discipline. We also emphasise that it is necessary to understand how social norms change over time, and how they are related to the development and expansion of the capitalist economic system, and how this influences related processes of consumerism and expanded needs that add to, rather than reduce, the practice of transactional sex, reflecting gendered abilities to access the fruits of development. It is essential that these insights are taken into account in the policy arena, and that development is not seen as a panacea for HIV prevention and the reduction of transactional sex, requiring a move away from the underlying poverty narrative.
We also discussed the implications for potential HIV prevention efforts of different economic approaches. Our analysis questions the transaction as the site of intervention, derived from economic approaches of which this is the focus, particularly in the light of the risk that women can face in engaging in economic activities, and evidence that suggests women from all social strata engage in transactional sex. The current popularity of this intervention is somewhat puzzling, especially as the evidence that microfinance reduces poverty is mixed and inconclusive (Duvendack et al. 2011), let alone expecting microfinance to somehow also tackle prevailing social norms around sex and exchange, or to provide women with more control over their sexual and reproductive lives. Further, this policy is based on a narrow and incomplete economic analysis, and to some extent represents an optimistic and unproven leap of faith. Whether policies such as these that attempt to work within the prevailing economic system (Bateman and Chang 2012) and that focus on small incentive changes without any attempts to engage with structural issues can be challenged, remains to be seen. However, alternative approaches derived from political economy and other economic approaches that are not based on the standard, individualistic, technical apparatus, offer a potential way forward. More research by political economists and those interested in the structural drivers of HIV is needed to better understand the historical, social and economic context of transactional sex, and how this practice evolves over time.