Saving souls, hurting bodies: Missions, health investments, and HIV prevalence in sub-Saharan Africa

Julia Cagé, Valeria Rueda 17 March 2018



From 2004 to 2016, the US government spent $70 billion to fight the HIV epidemic as part of the President’s Emergency Plan for AIDS Relief (PEPFAR).  In 2006, PEPFAR specified that 33% of all prevention funds (and two-thirds of funds for sexual transmission) would be earmarked for Abstinence Until Marriage programs (AUM). While many officials argue PEPFAR is a success, a number of recent studies find these AUM programmes to be a failure (Lo et al. 2016, Santelli et al. 2017). However, the HIV prevention curricula currently taught in schools in sub-Saharan Africa are often still largely shaped by Christian influence, tending to mainly focus on abstinence. HIV prevalence in African countries ranges from 1–25%, depending on countries and estimates; it thus seems crucial to investigate the effects of Christianity on the epidemic.

In a new working paper, we evaluate the long-term effects of missionary activity on HIV prevalence today (Cagé and Rueda 2017). On the one hand, the history of modern medicine in sub-Saharan Africa is closely linked to the development of missionary activity. Missionaries and militaries were among the first to invest in Western medicine in this part of the world (Vaughan 1991). According to the World Missionary Atlas (Dennis et al. 1903), there were 150 missionary physicians in Africa in 1925 and more than 235 nurses working with nearly 500 trained native nurses in 116 hospitals and 366 dispensaries.

On the other hand, health investments are not the only channel through which missionaries may have shaped HIV prevalence. One of the missions’ objectives, after all, was to convert the local population to Christianity, a religion that imposes a great number of explicit and implicit constraints on sexual behaviours (monogamy, abstinence outside and before marriage, frequent rejection of contraception, etc.). Given that conversion persisted (Nunn 2010), especially around Protestant missions, we can expect that missions affected the HIV epidemic through sexual behaviours and attitudes towards sex education. Furthermore, the Christian influence has strongly shaped prevention campaigns against HIV. Many of the initiatives to spread information about HIV on the field are conducted in churches (Mash and Mash 2013), and the HIV prevention curricula taught in schools emphasised abstinence and monogamy (Duflo et al. 2015, Dupas 2011).

We show that there is heterogeneity in the long-term effect of missions. On the one hand, regions close to missions in general tend to exhibit higher HIV prevalence. We find that less knowledge about condom use is a likely channel driving this effect. On the other hand, regions close to missions that specifically invested in health exhibit lower HIV prevalence today. Finally, the increased HIV prevalence around missions only seems to hold when considering Protestant missions.

To identify these effects, we overcome three main empirical challenges.

  • First, missions are located in favoured areas, often in proximity to the coast or large rivers. To address this concern, we restrict the analysis to regions located within a radius of 100km around missions.
  • Second, as urbanisation tends to be higher around missions, and HIV prevalence is a disease of urban areas, we show that our results are robust to focusing only on cities (and so not entirely driven by higher urbanisation).
  • Finally, when analysing missionary investments, there is the risk of capturing unobservable missionary characteristics if there is endogenous selection of missions into health services. We solve this problem by matching missions that invested in health to similar missions – based on observable geographical and historical characteristics – but that did not build hospitals or dispensaries.

Figure 1 Historical data and location of DHS cluster


We complement and improve the data collected for our previous research on the long-lasting effects of the printing press (Cagé and Rueda 2016). Our new database locates Protestant missionaries and their investments in schooling, health, or printing until 1903, as well as Catholic missionaries and their hospitals. The maps of Catholic missions are taken from the Vatican’s 1924 Atlas Hierarchicus. Figure 1 shows the map of historical missions with their health investments, as well as the locations of the Demographic and Health Survey (DHS) data from which we construct our measures of local HIV prevalence.

Heterogeneous effects

Our results show that missionary activity had heterogeneous effects on HIV depending on the mission denomination (Protestant versus Catholic), and depending on whether missions invested in health.

Figure 2 LOWESS plots: Distance to mission and HIV prevalence

Figure 2 shows the effect of the distance to any mission (top plot) and of the distance to a mission with a health investment (bottom plot) on HIV prevalence. There is a negative relationship between distance to a mission and HIV prevalence – in other words, the closer a region is to a mission, the higher the prevalence of the disease in this region. But the relationship goes the other way when we focus on missions with a health investment – HIV prevalence tends to steadily increase as we move away from a mission that invested in health.

To get a sense of the magnitude of the effects we obtain, we perform a simple thought experiment. Consider a country with 26 million inhabitants (our sample average) and assume that, everything else being equal, all the towns are 10km closer to a mission that they actually are. This would imply – absent any general equilibrium effects – a 24,000 increase in the number of HIV cases. In comparison, according to the most recent estimates, 26 million people are living with HIV in sub-Saharan Africa today.

The negative effect of proximity to a mission is driven by Protestant missions. If anything, proximity to Catholic missions seem to be associated with lower prevalence. In contrast, proximity to a missionary health investment decreases HIV prevalence regardless of the denomination.

Can urbanisation explain away the results?

How can one explain the negative effect of missionary proximity on HIV? While the Christian influence on norms is our favourite explanation, one may argue that a third factor may be driving our results, namely, urbanisation. HIV is indeed a disease of urbanisation, and it is well documented that missions brought about the development of large cities.

To check whether this is the case, we restrict our sample of analysis to urban areas located in regions with an average road density larger than the country's 75th percentile, and that are located close to the coast (<150km), a major determinant of missionary location. Then, in this sample of 126 urban areas, we regress HIV prevalence on the same controls as before, and on a binary variable equal to one if the urban area is located in a buffer around a historical mission settlement. Figure 3 presents the results. We can see that proximity to a Protestant mission is still associated with higher HIV prevalence today, whereas the effect is either nil or the opposite, but very small, for Catholic missions. Urbanisation is thus unlikely to be the only channel explaining the persistent effect of missionary activity on HIV prevalence.

Figure 3 HIV prevalence in urban areas close and far from missions

The role of conversion

The heterogeneity found between Catholic and Protestant missions strikes us as a surprise. After all, the most well-known opposition to anti-HIV policies that favour access to condoms was the Vatican's publication Family Values versus Safe Sex, of which the first chapter is entitled “The Catholic Church's Criticism of the Condom in AIDS-Prevention Programmes". Although surprising, the denominational heterogeneity in the results is consistent with an understanding of the phenomenon where conversion is a crucial channel.

Our results show indeed that Catholic missions were not nearly as successful as Protestants in their conversion enterprise. Figure 4 illustrates this difference in conversion rates. Similarly, we find that proximity to a mission is not statistically significantly associated to a change in HIV prevalence for towns where the majority of the population is Muslim.

Figure 4 Distribution of share of income across towns close to Protestant and Catholic missions

Sexual behaviours

Our results also show that people in towns close to any mission tend to be less abstinent before marriage, marry when they are older, and have more sexual partners per year. These are patterns in accordance with urban family structures. Therefore, efforts by faith-based organisations to encourage Christian family values as HIV prevention do not seem to be successful in these areas. This result is consistent with previous findings such as those in Dupas (2011), who shows that encouraging abstinence until marriage is inefficient at reducing risky sexual behaviours.

We also show that proximity to a mission that invested in health is associated with safer sexual behaviour. We observe lower declared use of sex workers’ services over a lifetime or during the year of the questionnaire. Women in these areas are also more likely to know where to find condoms.


There is a growing body of research on the effects of development programs. A book like Poor Economics (Banerjee and Duflo 2012) provides a plethora of examples from NGO-led initiatives that have evaluated their effects on diverse development outcomes, such as facilitating access to clean water, paying for vaccines, or distributing medication. During the colonial period, missions served a religious purpose, but also pursued an agenda that aimed to improve living conditions by increasing literacy, or investing in health. Their work is a tangle of projects that always bore these two objectives. Missionary work, in sum, is an ancestor of today’s development initiatives, but with a more pronounced religious flavour.

Our research shows that even ‘well-intentioned’ interventions can have unexpected effects in the long run. The growing body of research on missionary activity complements the development literature on programme evaluations by providing a very long-term perspective and uncovering complex effects.


Banerjee, A and E Duflo (2012), Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty, PublicAffairs.

Cagé, J and V Rueda (2016), "The long-term effects of the printing press in sub-Saharan Africa", American Economic Journal: Applied Economics 8(3): 69–99.

Cagé, J and V Rueda (2017), "Sex and the Mission: The Conflicting Effects of Early Christian Investments on the HIV Epidemic in sub-Saharan Africa", CEPR Discussion Paper 12192.

Dennis, J S, H P Beach and C H Fahs (eds) (1903), World Statistics of Christian Missions, Student Volunteer movement for foreign missions.

Duflo, E, P Dupas and M Kremer (2015), "Education, HIV, and early fertility: Experimental evidence from Kenya," American Economic Review 105(9): 2757–2797.

Dupas, P (2011), "Do teenagers respond to HIV risk information? Evidence from a field experiment in Kenya," American Economic Journal: Applied Economics 3(1): 1–34.

Lo, N C, A Lowe and E Bendavid (2016), "Abstinence funding was not associated with reductions in HIV risk behavior In sub-Saharan Africa," Health Affairs (Project Hope) 35(5): 856–863.

Mash, R and R Mash (2013), "Faith-based organisations and HIV prevention in Africa: A review," African Journal of Primary Health Care & Family Medicine 5(1).

Nunn, N (2010), "Religious conversion in Colonial Africa," American Economic Review Papers and Proceedings 100(2): 147–152.

Santelli, J S, L M Kantor, S A Grilo, I S Speizer, L D Lindberg, J Heitel, et al. (2017), "Abstinence-only-until-marriage: An updated review of US policies and programs and their impact," Journal of Adolescent Health 61(3): 273–280.

Vaughan, M (1991), Curing Their Ills: Colonial Power and African Illness, Stanford University Press.



Topics:  Development Economic history Health economics

Tags:  HIV prevalence, HIV, sub-Saharan Africa, health education, church, Christianity, Protestant Church, Catholic Church, health investments

Assistant Professor in Economics, Sciences Po; CEPR Research Fellow

Rokos Career Development Fellow, Pembroke College, Oxford University


CEPR Policy Research