Gender, ethnicity, and unequal opportunity in colonial Uganda: European influences, African realities, and the pitfalls of parish register data

by Michiel de Haas and Wourt Frankema (University of Wageningen)

The full article is published by The Economic History Review and it is available here


The Renaissance of African economic history in the past decade (see Austin and Broadberry 2014)  has opened up a vast body of qualitative and quantitative source materials.  Most of these materials were originally produced by European missionaries, merchants, travellers, or colonial officials, and thus reflect the biases (explicit or unspoken) of people who were alien to the societies to which their writings and statistics pertain. Moreover, many of the analytical concepts and empirical methods used by scholars to study these materials today, were originally designed for the study of European economic history. To prevent what Gareth Austin has described as ‘conceptual Eurocentrism’ in narratives of long-term African development, it is imperative to scrutinize and debate the applicability of concepts such as national income, real wages, human capital, social mobility, crime, or gender inequality in an African historical context.

In our article, we engage with a recent study by Meier zu Selhausen and Weisdorf to show how selection biases in, and Eurocentric interpretations of, parish registers have provoked an overly optimistic account of European influences on the educational and occupational opportunities of African men and women in Kampala, Uganda. We re-evaluate the link between colonial rule and gendered educational and occupational opportunities in Uganda along several lines. Here, we address two key points: 1) sample selection biases in African parish register data and 2) the flawed  juxtaposition of ‘African tradition’ and ‘European modernity’ to understand educational and occupational change during the colonial period.

Firstly, the use of parish registers by Meier zu Selhausen and Weisdorf illustrates clearly how data biases particular to the African context can result in doubtful conclusions. Parish registers have been widely used in European economic and demographic history, and are typically seen as fairly representative. However, the parish registers extracted from Kampala’s Anglican Namirembe Cathedral exhibit a strong bias towards the upper social classes: elites converted earlier, only the wealthiest and most well-connected African Christians opted for a ring marriage, and this particular Cathedral emerged as the elite church of early colonial Uganda.

To substantiate this issue quantitatively, we re-chart literacy and numeracy trends using microdata from the 1991 population census, made available by IPUMS. As Figure 1 shows, a representative selection of men and women from birth cohorts in the census data accumulated literacy and numeracy much slower than the population of brides and grooms in Namirembe Cathedral. While the Namirembe grooms achieve practically full literacy from the 1890s birth cohort onwards, and brides from the 1900s onwards, a random sample of Kampala-born men achieved similar rates only for cohorts born from the 1940s onwards, and women from the 1960s. That is: half a century later! Notably, Uganda’s female literacy take-off happened around independence, and the gap between men and women remained large throughout the colonial period.

There is ample reason to believe that similar (and/or potentially other) biases will be found among parish register samples in other African settings. Certainly, our study establishes that such registers, while a promising new source of microdata, should be interpreted with utmost caution.


Literacy rates and literacy gaps, per birth cohort

Secondly, we show that the dichotomy between African tradition and European modernity is flawed on both sides. The diffusion of Christianity and education in Buganda were not a testament to modernizing influences of Europeans, but rather the consequence of the initiative of the indigenous population of Buganda. Local elites pragmatically sought out political coalitions with European missionaries to further their own ends. Moreover, without the Africanization of the mission the diffusion of literacy and schooling would never have attained significant scale. Thus, the widely accepted idea that the uneven diffusion of missionary education can be explained by European supply factors rather than African demand (see a foundational paper by Nathan Nunn and a large body of subsequent studies) is not supported by the Ugandan case.

At the same time, the modernizing influence of Europeans was limited. Demand for African skilled labour in the colonial economy – which was predicated on rural cash crop production – was very limited, and its distribution was highly uneven, along lines of gender, race, ethnicity and location. Uganda’s economy remained overwhelmingly rural (colonial Kampala never harbored more than 1.5 per cent of Uganda’s population), and the skilled labour market was dominated by Asian and European expatriate minorities.

European influences also hardly benefited gender emancipation in Uganda. We revisit the seminal work of Ester Boserup and note, contrary to Meier zu Selhausen and Weisdorf, that she gives primacy to economic rather than cultural explanations for female labour market marginalization, and is skeptical about the emancipatory impact of European influences in Africa. In line with Boserup’s views, missionaries and colonial officials in Uganda often coalesced with indigenous patriarchal interests to domesticize women – a legacy that Ugandan women and society at large struggle with up until today.


To contact the authors:



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s