Impacts of long-lasting civil conflicts on education: Evidence from the 2014 Census of Myanmar

Hiroyuki Yamada, Midori Matsushima

Research output: Contribution to journalArticlepeer-review

Abstract

Geocoded conflict information was combined with the 2014 household census data to study the impact of long-lasting township-level internal conflicts on Myanmar's primary school attendance (i.e., the short-term impact) and years of education (i.e., the long-term impact). First, we constructed quasi-panel data for primary-level schooling to find consistently negative, but statistically insignificant, impacts of internal conflicts. The results are robust, even if incompleteness of census or migration are taken into account. In addition, the magnitudes of estimated impacts are much smaller than those of the findings from other countries. Second, we confirmed that conflict exposure from 6−10 years of age has a negative but insignificant impact on years of education. Gender differences in terms of negative impact are almost negligible. By carefully reviewing previous papers and the characteristics of Myanmar's conflicts until 2014 with respect to the mechanism of the negative effect of conflict on education, we argue that the small negative and statistically insignificant impact found in our analysis is due to the long-lasting and low-intensity nature of the conflicts, as well as the fact that schools and social services are provided by military forces. However, it is important to note that our analysis does not include data of the recent violence in Rakhine state.

Original languageEnglish
Article number101250
JournalJournal of Asian Economics
Volume71
DOIs
Publication statusAccepted/In press - 2020

Keywords

  • Civil conflict
  • Education
  • Myanmar
  • Population census

ASJC Scopus subject areas

  • Finance
  • Economics and Econometrics

Fingerprint Dive into the research topics of 'Impacts of long-lasting civil conflicts on education: Evidence from the 2014 Census of Myanmar'. Together they form a unique fingerprint.

Cite this