Share

BREAD Working Paper No. 256, March Gianmarco León (2010)

A study of post-conflict Peru finds that political conflict negatively affects future educational attainment.  Interestingly, the study finds that children in conflict situations can catch up in later educational attainment if they have already started school, but those who begin school after conflict has set in are permanently affected.

Peru was wracked by civil conflict from 1980-1993, a bloody period in which the rebel Partido Comunista del Perú-Sendero Luminoso (PCP-SL) clashed with the army.  Chafing against economic woes and growing inequalities in Peruvian society, this Maoist-inspired group declared war against the government.  The group’s power began to diminish only in 1992 when the head of the revolutionary army was captured.  Nearly 70,000 died in this era.  While these events are clearly significant for Peru’s development prospects, the impact of conflict on survivor welfare is an important question more broadly.

Previous studies found that conflict dramatically injures social welfare indicators, but that these indicators exhibit significant bounce back in the post-war period. 

However, these primarily cross-country studies have not dedicated sufficient attention to the impact of conflict on personal and household welfare.  In particular, the literature has not examined short and long term effects of such conflict.  For example, violence might lead to the demise of a household member, decreasing household income and nutrition and potentially child cognitive ability through the nutrition pathway.  Violence might also psychologically harm mothers or children.  It is important to examine the long run as well as the short run in order to determine if personal welfare bounces back after episodes of conflict as do overall welfare indicators.

León examines the impact of political violence on human capital accumulation.  He finds that political violence negatively impacts such accumulation.  He reports that “the average person exposed to political violence before starting school (in the early childhood, or in pre-school age) accumulates 0.21 less years of education as an adult.”  In the context of other studies, this figure is large.  Esther Duflo (2001) finds that a “massive” school construction project in Indonesia led to 0.12 to 0.19 more years of average schooling, while Akresh and de Walque (2010) find that children directly exposed to the Rwandan genocide lagged 0.5 years below average in primary schooling.  

In an interesting twist on this story, León finds that the degree to which violence affects educational outcomes depends on when children are exposed to violence. 

He reports that “Children who are affected by violence once they have already started their school life are able to catch-up, while the ones who experience violence before entering school are permanently affected” (20).  This provides support for a large literature on “critical-period programming” for children.

Additionally, León finds that short term effects of violence exceed long term effects, “showing that the persistence of the shock depends on the moment in life in which the child was exposed to violence” (2).  Another significant finding is that violence more negatively affects the educational attainment of women and native speakers (as compared to men and Spanish speakers, respectively).

León argues that there are supply and demand mechanisms driving these results. 

On one hand, violence decreased the supply of schooling in Peruvian districts where a teacher was killed and school entrance was therefore delayed.  León reports that approximately three percent of human rights violations were perpetrated against teachers, and that “it was not an easy task to replace a teacher in a violent area” (12).  On the other hand, violence negatively impacted maternal and child health, decreasing the demand for the schooling of those children.  After an examination of the data, León concludes that the teacher pathway is important in the short run but not the long run, while maternal health shocks affecting child health and attainment are important in the long run.

Methodologically, he compares the number of human rights violations in Peru’s conflict era with 1993 and 2007 census data.  The human rights dataset was assembled by the Peruvian Truth and Reconciliation Commission and contains “detailed records of every human rights violation reported during the period of civil violence in Perú” (7).  León analyzes the incidence of illegal detentions, kidnappings, murder, extrajudiciary executions, torture, and rape in particular.  He compares this district-level “violence shock” to the educational outcomes of children born in those districts.  Limitations of this methodology include sample composition (human rights violations were reported voluntarily to the Commission and therefore are a non-random sample), and the fact that the census data does not capture migration effects.  However, León’s robustness checks suggest that these limitations do not significantly affect the results.  In any case, these two problems would bias the estimated effect of violence downwards, making the model more conservative.

León concludes with a strong policy recommendation.  He writes that “Overall, the results in this paper reinforce the idea that shocks during the early stages of one’s life have long term irreversible consequences on human welfare.  Further, relief efforts in the case of violence shocks should be targeted first to pregnant mothers and young children, and then to children in the early stages of their schooling cycle if we want to minimize the long term welfare losses for society” (23-24).