UC Berkeley

How does exposure to genocide affect nuptiality and fertility among the surviving population? Genocides in Rwanda and in Bosnia-Herzegovina in the early 1990s caused high levels of population displacement, trauma, and death, along with a dramatic decline in the standard of living. In Rwanda, genocide also reduced the sex ratio of the marriage-aged population, while in Bosnia, despite the high proportion of male casualties, the overall sex ratio of the marriage-age population did not decline substantially. Contrary to the expected delay in first marriage throughout the process of demographic transition, there was clear evidence of a post-genocide marriage boom among younger cohorts in Rwanda. Meanwhile in Bosnia--a country that had already achieved the demographic transition and fairly steady marriage rates more than a decade before the war started--first marriage rates peaked slightly during and after the war and then continued to decrease during the decade after the war ended, never returning to prewar levels. Patterns in fertility echoed this trend, with a small baby boom in Rwanda contrasted with a fall in birth rates in Bosnia.

Drawing from a multi-method, multi-site study involving data analysis from nationally representative household surveys, 117 qualitative life-history interviews and 34 interviews with key informants in both countries, I assess the relative importance of several factors on marriage and fertility trends during and immediately after genocide in both countries. By examining the Bosnian and Rwandan cases in tandem, I situate demographic changes due to genocide in a broader social and historical framework.

It is difficult to compare the demographic effects of war and genocide in two countries that had widely divergent pre-war nuptial and reproductive contexts. I explore four broad sets of causal factors linking genocide with demographic outcomes: (1) involuntary factors, such as mortality, separation of partners, and survival bias; (2) material and economic factors, including loss of income, changes in employment, and `transactional marriage' during acute periods of crisis; (3) sex ratio and gender role factors, which encompass the change in the marriage market due to differential mortality by sex, changing gender roles during and after wartime, and norms of widowhood remarriage; and (4) psychosocial factors, including the duration and severity of conflict, the effects of rape and post-traumatic stress disorder [PTSD] on the surviving population, and the role of political recovery and social cohesion in demographic decisions and outcomes.

Whereas many journalists emphasized ethnic hatreds as the motivation behind both genocides, the instrumentalist view of ethnic conflict--which claims that ethnic divisions are exploited by elites during moments of economic and political crisis--has emerged as a powerful counter-narrative in both Bosnia and Rwanda. In areas where it is possible to measure ethnic differences my findings tend to support the instrumentalist view--namely, that economic and situational factors were more salient than ethnic differences in determining the demographic response to genocide.

Broadly speaking, the mortality effects of genocide dominated in Rwanda, and this helped fuel an increase in marriages and births after war. Meanwhile in Bosnia the economic effects of genocide and the accompanying transition to a market economy tended to dominate the demographic response to genocide, suppressing marriage and fertility rates during and after the war. To classify the dominant character of these genocides in this way is not to ignore the massive economic collapse in Rwanda, nor the extraordinary death toll in Bosnia, but rather to suggest the relative primacy of economic and mortality effects as mediating channels through which genocide affected demographic outcomes for the majority of the population in each country.

I postulate the need for a typology of the demographic effects of crisis. Famine, economic collapse, political instability, natural disaster, and armed conflict each affect population dynamics in different ways. Given the variation and complexity between the Bosnian and Rwandan genocides, it would be difficult to extrapolate directly from these two cases. Yet, when considered in tandem, the Bosnian and Rwandan case studies suggest a need to consider several important axes of variation--factors such as the intensity and duration of the crisis, economic effects, differential age structure effects, displacement, mortality, background conditions, and post-crisis political and social stability--when predicting the effects of future crises on nuptial and reproductive trends.