This thesis consists of three empirical studies which analyse key socioeconomic barriers to a
country’s prospects of long-term development and growth. Using Colombia as a case study, the
following analyses focus on deepening our understanding of crucial issues, ranging from financial
literacy and economic behaviour, to the effect of conflict-related violence on social preferences
and youth’s wellbeing.
The first empirical study assesses financial literacy levels amongst Colombian consumers and its
role in promoting optimal economic behaviour. Following the approach developed by Lusardi
and Mitchell (2008; 2009; 2011a), this chapter examines individuals’ understanding of three financial
concepts, including interest rate compounding, inflation rates and risk diversification.
Moreover, it relies on two novel instruments, namely local development funds and the number
of schools at the locality level, to control for the potential endogeneity of financial literacy. The
results indicate that financial illiteracy is widespread across our sample, with only 52% of respondents
understanding interest rate compounding and less than half of individuals grasping
both inflation rates and risk diversification. Most importantly, the findings show a positive relationship
between financial illiteracy and welfare-reducing economic practices. Specifically, those
who are financially savvy are not only less likely to rely on informal borrowing methods but are
also significantly less likely to incur debt as a means to make ends meet.
The second empirical study examines the impact of Colombia’s 58-year-long armed conflict on
individual-level social outcomes, including political polarisation and attitudes towards the use
of violence, vigilantism and iron-fist policies. To address potential sources of endogeneity, the
identification strategy relies on a difference-in-differences model that exploits variation in the
geographical and temporal distribution of municipality-level violence during 2013-2016. The
findings show that conflict exposure is a key determinant of social preferences. In particular,
respondents living in areas highly exposed to violence are significantly less likely to support the
use of violence, relative to their counterparts in lower exposure municipalities. Interestingly,
whilst individuals living in conflict-ridden areas are significantly less likely to support iron-fist
policies that erode the rule of law, high exposure to threats is also associated with a greater
likelihood of individual-level political polarisation.
Building on the analysis of the previous chapter, the final study examines the impact of conflict-related
violence on the probability of risk behaviour amongst 9-16-year-old children and adolescents.
While the analysis focuses predominantly on alcohol use, it also examines the effect of
violence on additional measures of externalising behaviours. Using a child fixed effects model,
the results show that changes in exposure to threats and forced displacement increase the probability
of alcohol use, particularly amongst 13-16-year-old adolescents. By contrast, we find
no significant effects of conflict exposure amongst 9-12-year-old children. This evidence is consistent
with research from developmental neuroscience which highlights mid-adolescence as a
critical period for the emergence of risk behaviour. Additionally, we find no gender differences
in the impact of violence, with both older female and male adolescents experiencing a significant
increase in their likelihood of alcohol use. Finally, we also find a significant effect of changes in
conflict exposure on our additional measures of externalising behaviours and index scores that
summarise adolescents’ maladaptive outcomes.
