Friday, 13 April, 2018 | 11:00 | Defense

Dejan Kovač: "Essays on Conflicts and Human Capital Accumulation"

Dissertation Committee:
Randall K. Filer (chair)
Štěpán Jurajda (local chair)
Jacob N. Shapiro



In the first chapter of the dissertation, parental mortality is associated with a range of negative child outcomes. This paper studies the effect of paternal mortality on children's health and schooling outcomes using the universe of veterans' children born in Croatia, and all of the paternal deaths and injuries resulting from the 1991-1995 Croatian-Serbian war.  Using linked administrative data, I find large negative effects of paternal death on high-school GPA, school absences, behaviour problems, and hospitalisations.  I address potentially non-random selection into paternal death by using within-military unit differences in the extent of injury or death, essentially assuming that the members of a military unit all had similar probabilities of being killed or injured because they fought in the same battles.  I am also able to shed light on an important mechanism underlying the estimated effects. Surviving spouses of those killed or injured were well compensated, so that the death of a father did not have a negative effect on household incomes. I find that a death or injury that occurred during the in-utero period has much larger effects than a death or injury in early childhood, suggesting that much of the negative effect is due to maternal stress

In the second chapter, we answer the following research question. How important is intergenerational transmission of political values for sustaining ethnic divisions and nationalism? We implement a novel empirical strategy for identifying and studying a strong form of nationalism---the willingness to fight and die in a war for national independence---using name choices corresponding to previous war leaders. Based on data on almost half a million soldiers, we first show that having been given a first name that is synonymous with the leader(s) of the Croatian state during World War II predicts volunteering for service in the 1991-1995 Croatian war of independence and dying during the conflict. Next, we use the universe of Croatian birth certificates and the information about nationalism conveyed by first names to suggests that in ex-Yugoslav Croatia, nationalism was on a continuous rise starting in the 1970s, consistent with Fearon and Laitin (2003), and that its rise was curbed in areas where concentration camps were located during WWII. We also link the nationalist values we proxy using first name choices to right-wing voting behavior in 2015, 20 years after the war. Our evidence on intergenerational transmission of nationalism is consistent with the trade-off between within-family and society-wide transmission channels of cultural values proposed by Bisin and Verdier (2001) and it suggests a mechanism that sustains elevated parochial altruism across generations.

In the third chapter, we ask whether the long-term health costs of combat experience are larger for soldiers exposed to higher risk of dying, i.e., to higher killed-in-action (KIA) rates of their units. We study the overall mortality effects and also focus on the suicide effects over a 20-year post-war period using data on all veterans of the 1991-1995 Croatian war for independence. The length of combat service predicts suicide risks, as does the exposure to psychological shocks such as the maximum experienced monthly unit-specific KIA rate. In the second part of the analysis we ask how war-related suicides affect the children of veterans and find significant negative effects on GPA, school absence and school behaviour. Comparing the children of soldiers who served within the same unit, we approximate the amount of psychological shock that each soldier experienced. Subsequently, we analyse the connection between selection into suicides and children's outcomes by comparing two groups of children: those whose fathers died from natural causes and those whose fathers committed suicide, in the post-war period. We find no differences in children's GPA across the two groups,. We find a significant difference in children's performance on high-stakes exams, where outcomes can determine future life trajectories. Remarkably, the children of veterans who committed suicide have lower scores on high school state exams than the children of veterans who died of natural causes. These findings are evidence how war costs are transmitted from one generation to another and how policymakers should internalise these unobserved costs into future policymaking.


Full Text: "Essays on Conflicts and Human Capital Accumulation" by Dejan Kovač