Social capital and conflict: impact and implications

This thesis explores the relationship between social capital and conflict in two different contexts, by answering the following two questions: How does exposure to violence affect social capital in urban Maharashtra, India? How does returning home affect social capital amongst internally displaced persons and returnees from Nahr el Bared camp in North Lebanon? This thesis then goes on to look at the labour market implications of returning home to Nahr el Bared camp, exploring the role of social capital (amongst other mechanisms) in this relationship. The following paragraphs are abstracts from the three empirical chapters that address these questions. The first empirical chapter explores the relationship between exposure to riots and social capital in urban Maharashtra. We exploit a panel dataset collected by the authors and apply a random effects model with lagged covariates to estimate an exogenous relationship between neighbourhood exposure to riots and four forms of social capital: membership in a group or organisation, trust in neighbours, participation in community discussions and participation in community festival preparations. Consistent with Bellows and Miguel's study of conflict and social capital (2009), we find that households living in neighbourhoods that experienced a riot are more likely to be members of groups and organisations. On the other hand, we find that these households are less likely to join community discussions, which lends more to the hypothesis of fragmented post-conflict societies with a damaged social fabric (Colletta and Cullen, 2000). We explore various mechanisms behind these results and find that the increased membership in organisations is greatest in diverse neighbourhoods that have not experienced recent changes in composition. However, riots reduce trust and the likelihood of participation in fragmented and polarised riot-affected neighbourhoods. Riots also decrease participation in festival preparations in neighbourhoods where out-migration has been low. Our analysis suggests that individuals and households instrumentally use social capital to their advantage, a type of insurance to protect against potential communal violence in the future. However, riots can have adverse affects on different forms of social capital that go beyond the surface level of social networking to feelings of trust and sense of community. The second empirical chapter studies the effect of returning home after conflict induced displacement on social capital, compared to remaining displaced. I have collected a household survey of displaced Palestinians from a refugee camp in Lebanon, and this chapter assesses the impact of return on the different dimensions of social capital based on a diverse and rich set of questions. An instrumental variable is used to model the return decision in one part of the camp, and the exogenous nature of return is exploited in another section of the camp. Results show that return can improve social capital if households return within one year of the war ending and with their friends and family. If households have been displaced for too long, then social capital is decreased upon returning home. This indicates that social capital is not simply carried over from displacement to return, but is rebuilt in a process that takes time and effort. The third and final empirical chapter studies the effect of returning home on labour market outcomes. Theoretically the effect of return is ambiguous, depending on changes in both the demand and supply of labour. I empirically study the effect of return on four labour market outcomes: participation in the labour force, working, wages and number of days worked. I analyse a dataset of individuals originally from Nahr el-Bared camp in North Lebanon, displaced within Lebanon after a war in 2007 between the Lebanese army and Fatah al-Islam. I use an instrumental variable and exploit the exogenous nature of the return process in order to estimate a causal effect of return. The results show that return increases the likelihood of working by 117 percentage points. This effect is greatest for those who have returned within two years, reaping the benefits of greater aggregate demand as the market increases. Women returnees are more likely to be working compared to the displaced, but there is no difference in employment between men who have been displaced and those who have returned. This could be because women possess skills that are adaptable in labour markets, working in cottage type industries from home, as opposed to the more specialised skills that men tend to possess.