The impact of extreme weather events on education

Abstract

This paper provides new evidence on the long- and medium-term impact of extreme weather events on education. Our focus is on Mongolia, where two extremely severe winters caused mass livestock mortality. We use household panel data with information on households’ preshock location, combined with historic district-level livestock census data and climate data. Our econometric strategy exploits exogenous variation in shock intensity across space and time, using a difference-in-differences approach. Results indicate that individuals who experience the shock while of schooling age and living in severely affected districts are significantly less likely to complete mandatory education, both in the long and medium terms. The effects are driven by individuals from herding households, while no significant effects are found for individuals from nonherding households. This finding renders it unlikely that extreme winters affect education through school closures during extreme climatic conditions, to which all children were exposed. Moreover, there is no evidence for a differential impact of extreme weather events by gender. This suggests that the effects are not mainly channeled through increased child labor in herding but rather they are related to reductions in household income.

This is a preview of subscription content, access via your institution.

Fig. 1

Notes

  1. 1.

    Related fields of research focus on the impact of extreme weather events on household income, anthropometric outcomes, health, and child labor (see Baez et al. 2010 for a comprehensive review).

  2. 2.

    Since the early 1990s, several reforms were implemented in the educational sector, including lowering the age of school entry and increasing the number of grades both in primary and in secondary school.

  3. 3.

    The livestock census data available to us includes both the overall number of adult animals that died as well as the subgroup of adult animals that died from disease. Livestock losses caused by disease account for about 5% of total losses in nondzud years and about 1.5% of total losses in dzud years. While the number of disease-related losses generally declined after 1993, there were peaks of disease-related deaths around dzud years. This suggests that livestock disease could be partly endogenous to dzud, when weakened animals become more susceptible to infections. For this reason, in all analyses presented in the following, we use the total number of livestock death (caused by natural disaster, disease, and other factors).

  4. 4.

    See the notes below Table 1 for information on the data sources and the exact definition of variables.

  5. 5.

    The summary of the 1999–2002 triple dzud relies on various reports, emergency appeals, and operation updates by the International Federation of Red Cross and Red Crescent Societies (IFRC). The summary of the 2009/2010 dzud builds on reports from the European Commission (2010), IFRC and MRCS (2010), the United Nations Mongolia Country Team (2010a), and Shinoda and Nandintsetseg (2015).

  6. 6.

    Formal insurance markets were not well developed in rural Mongolia when the considered dzuds occurred. Thus, apart from emergency aid provided by the government and international agencies on an ad hoc basis, herding households were largely left to cope with dzud on their own.

  7. 7.

    Unfortunately, no data is available on school closures.

  8. 8.

    The province is the top level of Mongolia’s administrative structure. Each province is subdivided into several districts, which are further subdivided in subdistricts (bags). The 49 districts included in our survey have an average size of 4865 km2 and an average population of 1002 households.

  9. 9.

    A PSU is the smallest population unit within Mongolia’s administrative division. PSU are entirely virtual and only used for administrative purposes. On average, a PSU comprises about 34, 58, and 81 households in rural areas, district centers, and provincial centers, respectively.

  10. 10.

    Within the sample of 777 individuals, 33 individuals changed the district of residence between 2009 and the survey interview of wave three. To assess the impact of using the exact location of residence just before the shock unfolded, we estimated regressions based on individuals’ location at the time of the survey, instead of their location in 2009. While all main results are confirmed, regression coefficients are slightly smaller in magnitude (available upon request). This underlines that postshock migration, if not accounted for, would lead to biased estimates.

  11. 11.

    Out of a starting regression sample of 2251 individuals (see Section 6 for details on sample composition), 419 individuals (18.6%) are excluded because they moved in the current district in the year 2000 or thereafter. We conducted balance tests comparing individual and household characteristics between the included and the excluded samples. As Table 9 in the Appendix shows, neither the outcome variable nor the household-level controls differ significantly across the two samples. Of the individual-level controls, it appears that women, older individuals, and individuals who are the spouse of the head are significantly more likely to be excluded from the sample. Seventy-eight percent of individuals excluded from the sample are married. Among those that report the year of marriage, 64% married in 2000 or thereafter. Hence, the majority of the excluded cases can be explained by marriage-related migration taking place between 2000 and 2012, with the bride joining the household of her husband. Because this kind of migration is not related to the dzud, we consider it unlikely that our results on the impact of the dzud are biased.

  12. 12.

    Adverse climatic conditions started in summer and autumn 1999, while mass livestock deaths occurred between January and May 2000 (also see Fig. 1). The same pattern holds for each dzud winter analyzed here. For this reason, we use livestock mortality of the following year when constructing the livestock-based dzud index.

  13. 13.

    By definition, individuals in this sample reached age 15 by 2008. Hence, we can exclude the possibility that the 2009/2010 dzud influenced the completion of mandatory schooling for them. Individuals in this sample started school before 2005 and hence followed the schooling system in force before 2005, which entailed 8 years of mandatory education (from 8 through 15 years of age).

  14. 14.

    We compared education levels measured in our survey with another (nationally representative) household survey implemented in Mongolia, the Multiple Cluster Indicator Survey (MICS) of 2010. According to the MICS, 76.6% of head of households in Mongolia completed at least basic education. In the first wave (implemented 2012/2013) of the Coping with Shocks in Mongolia Survey, which our analysis builds on, this figure is 73.3%. We take this as supportive evidence of the quality of our data.

  15. 15.

    Year of birth fixed effects also absorb differences across cohorts in the exposure to the transition period.

  16. 16.

    The number of livestock a household owned in 2009 was recorded retrospectively during the first wave interview in 2012/2013. Three considerations based on empirical evidence make us confident that the retrospectively recorded information on past livestock holdings is reliable. First, anthropological studies on Mongolia stress the importance of livestock holdings for the social standing of households. For instance, there are specific terms in the Mongolian language to classify herders with different livestock holdings (<100 heads; 100–200; 200–500; 500–1000; and >1000 heads) (Murphy 2011). Therefore, it is not surprising that our survey enumerators did not observe difficulties among respondents to recall the size of their herd in the past. Second, households are asked about their livestock holdings in 2009 twice, in the first panel wave and again in the third panel wave. The coefficient of correlation for livestock holdings in 2009 recorded in the first and third wave is 0.79. Third, we tested if dzud intensity affected households’ ability to report pre-dzud livestock holdings. We regressed the number of livestock a household owned in 2009 on various district-level measures of intensity of the 2009/2010 dzud. Results (available upon request) show that none of the district-level shock measures is significantly correlated with the number of livestock households reported to own in 2009. Hence, this exercise does not give reason for concern that dzud experience is correlated with the ability to recall.

  17. 17.

    Similar results are obtained for shock measures based on livestock mortality, snow depth, and NDVI when we account separately for each of the three dzud winters occurring in 1999/2000, 2000/2001, and 2001/2002 (results available upon request).

  18. 18.

    Given space constraints, Table 5 only presents results for shock intensity measures defined as above the 75th percentile threshold. Measuring shock intensity with the 85th percentile threshold or continuous measures brought very similar results for both panel A and panel B (results available upon request).

  19. 19.

    The estimated effects of the 2009/2010 dzud and the 1999–2002 triple dzud are not directly comparable in magnitude, for a number of reasons: The length and intensity of each shock was different, while also different climatic conditions were at play. In addition, we estimate the effect of the 1999–2002 triple dzud for children exposed in both primary and secondary school age, while the timing of the data collection constrains us to estimate the effect of the 2009/2010 dzud only for children exposed in secondary school age.

References

  1. Alderman H, Hoddinott J, Kinsey B (2006) Long term consequences of early childhood malnutrition. Oxf Econ Pap 58(3):450–474

    Article  Google Scholar 

  2. Baez JE, de la Fuente A, Santos I (2010) Do natural disasters affect human capital? An assessment based on existing empirical evidence. IZA Discussion Papers 5164

  3. Barro R, Lee J-W (2013) A new data set of educational attainment in the world, 1950–2010. J Dev Econ 104(C):184–198

    Article  Google Scholar 

  4. Batima P (2006) Climate change vulnerability and adaptation in the livestock sector of mongolia. International Start Secretariat, Washington

  5. Björkman-Nyqvist M (2013) Income shocks and gender gaps in education: evidence from Uganda. J Dev Econ 105:237–253. doi:10.1016/j.jdeveco.2013.07.013

    Article  Google Scholar 

  6. Cunha F, Heckman J (2007) The technology of skill formation. Am Econ Rev 97(2):31–47. doi:10.1257/aer.97.2.31

    Article  Google Scholar 

  7. Dairii A, Suruga T (2006) Economic returns to schooling in transition: a case of Mongolia. GSICS Working Paper Series 9

  8. De Vreyer P, Guilbert N, Mesple-Somps S (2015) Impact of natural disasters on education outcomes: evidence from the 1987–89 locust plague in Mali. J Afr Econ 24(1):57–100

    Article  Google Scholar 

  9. Deuchert E, Felfe C (2015) The tempest: short- and long-term consequences of a natural disaster for children’s development. Eur Econ Rev 80:280–294. doi:10.1016/j.euroecorev.2015.09.004

    Article  Google Scholar 

  10. Engel J, Prizzon A (2014) From decline to recovery: post-primary education in Mongolia. Overseas Development Institute, London

    Google Scholar 

  11. European Commission (2010) Commission decision on the financing of humanitarian actions in Mongolia from the general budget of the European Union (ECHO/MNG/BUD/2010/01000). European Commission, Brussels

    Google Scholar 

  12. Fernandez-Gimenez ME (1999) Sustaining the steppes: a geographical history of pastoral land use in Mongolia. Geogr Rev 89(3):315–342

    Article  Google Scholar 

  13. Goodman J (2014) Flaking out: student absences and snow days as disruptions of instructional time. NBER Working Paper 20221

  14. Groppo V, Kraehnert K (2016) Extreme weather events and child height: evidence from Mongolia. World Dev 86:59–78

    Article  Google Scholar 

  15. Guarcello L, Mealli F, Rosati FC (2010) Household vulnerability and child labor: the effect of shocks, credit rationing, and insurance. J Popul Econ 23(1):169–198. doi:10.1007/s00148-008-0233-4

    Article  Google Scholar 

  16. Hansen B (2011) School year length and student performance: quasi-experimental evidence. Unpublished manuscript, University of Oregon

  17. Horton S, Steckel RH (2013) Malnutrition: global economic losses attributable to malnutrition 1900–2000 and projections to 2050. In: Lomborg B (ed) How much have global problems cost the world? A scorecard from 1900 to 2050. Cambridge University Press, Cambridge, pp 247–272

    Chapter  Google Scholar 

  18. International Federation of Red Cross and Red Crescent Societies (IFRC), Mongolian Red Cross Society (MRCS) (2010) Rapid assessment of dzud situation in Mongolia (January 18–January 26, 2010): summary report. IFRC and MRCS, Ulan Bator

    Google Scholar 

  19. Jacoby HG, Skoufias E (1997) Risk, financial markets, and human capital in a developing country. Rev Econ Stud 64(3):311–335

    Article  Google Scholar 

  20. Jensen R (2000) Agricultural volatility and investments in children. Am Econ Rev Pap Proc 90(2):399–404

    Article  Google Scholar 

  21. Justino P, Leone M, Salardi P (2014) Short- and long-term impact of violence on education: the case of Timor Leste. World Bank Econ Rev 28(2):320–353. doi:10.1093/wber/lht007

    Article  Google Scholar 

  22. Lawrie J, Dandii O (2010) Report on the 2009–10 dzud disaster impact on schools, kindergartens, children and teachers in Mongolia. Save the Children Japan, Ulaanbaatar

    Google Scholar 

  23. León G (2012) Civil conflict and human capital accumulation: the long-term effects of political violence in Peru. J Hum Resour 47(4):991–1021

    Google Scholar 

  24. Maccini S, Yang D (2009) Under the weather: health, schooling, and economic consequences of early-life rainfall. Am Econ Rev 99(3):1006–1026

    Article  Google Scholar 

  25. Marcotte DE (2007) Schooling and test scores: a mother-natural experiment. Econ Educ Rev 26(5):629–640

    Article  Google Scholar 

  26. Marcotte DE, Hemelt SW (2008) Unscheduled school closings and student performance. Educ Finan Policy 3(3):316–338

    Article  Google Scholar 

  27. Middleton N, Rueff H, Sternberg T, Batbuyan B, Thomas D (2015) Explaining spatial variations in climate hazard impacts in western Mongolia. Landsc Ecol 30:91–107

    Article  Google Scholar 

  28. Miller RT, Murnane RJ, Willett JB (2008) Do teacher absences impact student achievement? Longitudinal evidence from one urban school district. Educ Evavl Policy Anal 30(2):181–200

    Article  Google Scholar 

  29. Ministry of Education Culture and Science of Mongolia (2006) Master plan to develop education of Mongolia in 2006–2015. Government of Mongolia, Ulaanbaatar

    Google Scholar 

  30. Murphy DJ (2011) Going on Otor: disaster, mobility, and the political ecology of vulnerability in Uguumur, Mongolia. PhD dissertation, University of Kentucky

  31. Murray V et al (2012) Case studies. In: Field CB, Barros V, Stocker TF, Qin D, Dokken DJ, Ebi KL, Mastrandrea MD, Mach KJ, Plattner G-K, Allen SK, Tignor M, Midgley PM (eds) Managing the risks of extreme events and disasters to advance climate change adaptation: a special report of working groups I and II of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, pp 487–542

    Chapter  Google Scholar 

  32. National Statistical Office of Mongolia (2003) Mongolian national statistical yearbook 2002. NSO, Ulaanbaatar

    Google Scholar 

  33. National Statistical Office of Mongolia (2013) Mongolian statistical yearbook 2012. NSO, Ulaanbaatar

    Google Scholar 

  34. National Statistics Office, UNICEF (2011) Multiple indicator cluster survey 2010: summary report. National Statistics Office, Ulan Bator

    Google Scholar 

  35. Palat Rao M et al (2015) Dzuds, droughts, and livestock mortality in Mongolia. Environ Res Lett 10:1–12. doi:10.1088/1748-9326/10/7/074012

    Google Scholar 

  36. Pastore F (2010) Returns to education of young people in Mongolia. Post-Communist Econ 22(2):247–265

    Article  Google Scholar 

  37. Rosales MF (2014) Impact of early life shocks on human capital formation: El Niño floods in Ecuador. IDB Working Paper Series 503

  38. Rouse JW, Haas RH, Scheel JA, Deering DW (1974) Monitoring vegetation systems in the Great Plains with ERTS. In: Freden SC, Mercanti EP, Becker MA (eds) Third earth resources technology satellite-1 symposium-volume: I technical presentations. NASA, Washington, pp 309–317

    Google Scholar 

  39. Seneviratne SI et al (2012) Changes in climate extremes and their impacts on the natural physical environment. In: Field CB, Barros V, Stocker TF, Qin D, Dokken DJ, Ebi KL, Mastrandrea MD, Mach KJ, Plattner G-K, Allen SK, Tignor M, Midgley PM (eds) Managing the risks of extreme events and disasters to advance climate change adaptation: a special report of working groups I and II of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, pp 109–230

    Chapter  Google Scholar 

  40. Shah M, Steinberg BM (2017) Drought of opportunities: contemporaneous and long term impacts of rainfall shocks on human capital. J Polit Econ (forthcoming)

  41. Shemyakina O (2011) The effect of armed conflict on accumulation of education: results from Tajikistan. J Dev Econ 95(2):186–200

    Article  Google Scholar 

  42. Shinoda M, Nandintsetseg B (eds) (2015) Climate change and hazards in Mongolia. Nagoya University, Nagoya

    Google Scholar 

  43. Skees J, Enkh-Amgalan A (2002) Examining the feasibility of livestock insurance in Mongolia. World Bank Policy Research Paper 2886

  44. Steiner-Khamsi G (2007) Mongolia country case study: country profile commissioned for the EFA global monitoring report 2008, education for all by 2015—will we make it? UNESCO, Paris

    Google Scholar 

  45. Sternberg T (2010) Unravelling Mongolia’s extreme winter disaster of 2010. Nomadic Peoples 14(1):72–86

    Article  Google Scholar 

  46. Tachiiri K, Shinoda M, Klinkenberg B, Morinaga Y (2008) Assessing Mongolian snow disaster risk using livestock and satellite data. J Arid Environ 72(12):2251–2263

    Article  Google Scholar 

  47. Townsend RM (1994) Risk and insurance in village India. Econometrica 62(3):539–591

    Article  Google Scholar 

  48. Udry C (1994) Risk and insurance in a rural credit market: an empirical investigation in northern Nigeria. Rev Econ Stud 61(3):495–526

    Article  Google Scholar 

  49. UNDP, National Emergency Management Agency (NEMA) (2010) Dzud national report 2009–2010. UNDP and NEMA, Ulaanbaatar

    Google Scholar 

  50. United Nations (2000) Mongolia: United Nations Inter-Agency appeal for Mongolia “DZUD 2000”—an evolving disaster. UN Disaster Management Team, New York

    Google Scholar 

  51. United Nations Mongolia Country Team (2010a) Mongolia 2010: dzud appeal. United Nations, Ulan Bator

    Google Scholar 

  52. United Nations Mongolia Country Team (2010b) Situation report no. 1: severe winter weather. United Nations, Ulan Bator

    Google Scholar 

  53. Valente C (2014) Education and civil conflict in Nepal. World Bank Econ Rev 28(2):354–383. doi:10.1093/wber/lht014

    Article  Google Scholar 

  54. Verwimp P, Van Bavel J (2014) Schooling, violent conflict, and gender in Burundi. World Bank Econ Rev 28(2):384–411. doi:10.1093/wber/lht010

    Article  Google Scholar 

  55. World Bank (2006) Public financing of education. Equity and efficiency implications. World Bank, Washington

    Google Scholar 

  56. World Bank (2010) World development report 2010: development and climate change. World Bank, Washington

    Google Scholar 

  57. World Bank (2015) World development indicators: education statistics (last accessed 8th July 2015). World Bank, Washington

    Book  Google Scholar 

  58. Zimmerman FJ, Carter MR (2003) Asset smoothing, consumption smoothing and the reproduction of inequality under risk and subsistence constraints. J Dev Econ 71(2):233–260

    Article  Google Scholar 

Download references

Acknowledgments

We gratefully acknowledge the helpful suggestions and guidance of three anonymous reviewers. We also thank Veronika Bertram-Hümmer, Adam Lederer, and Francesco Pastore for helpful comments. Bayarkhuu Chinzorigt, Jan Eberle, Maximilian Huppertz, Carlotta Nani, Johannes Matzat, and Ramona Schachner provided excellent research assistance. The paper also benefited from comments received at the NCDE 2015 in Copenhagen, the EEA 2015 in Mannheim, the 2015 Annual Conference of Verein für Socialpolitik (VfS) in Münster, the 2015 Annual Conference of the VfS Research Group on Development Economics in Kiel, and a workshop at DIW Berlin. We are grateful to our Mongolian partner, the National Statistical Office of Mongolia, for the fruitful cooperation in collecting household survey data. The responsibility for the content of this paper lies solely with the authors.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Kati Kraehnert.

Ethics declarations

Funding

The research was funded by the German Federal Ministry of Education and Research, funding line Economics of Climate Change, research grant 01LA1126A.

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Responsible editor: Erdal Tekin

Appendix

Appendix

Fig. 2
figure2

Map of Mongolia, showing the location of the survey area

Fig. 3
figure3

Spatial variation in intensity of the 1999–2002 triple dzud across survey districts. Sources: NESDIS STAR VHP (NDVI), ERA-Interim dataset (snow), and Mongolia Livestock Census

Fig. 4
figure4

Spatial variation in intensity of the 2009/2010 dzud across survey districts. Sources: ERA-Interim (snow and temperature) and Mongolia Livestock Census

Fig. 5
figure5

Overview of the identification strategy. Note: The figures show the age of sample individuals in each year between 1976 and 2012 (panel A) and between 1989 and 2014 (panel B). Each survey wave is collected over a 12-month period, from June to May of the following year. For simplicity, only the first year of each survey wave is shown

Fig. 6
figure6

Preshock education, 2009/2010 dzud. Sources: ERA-Interim (snow and temperature) and Mongolia Livestock Census

Table 9 Balance tests for individuals in the included and excluded samples (dzud 1999–2002)
Table 10 Pre-dzud average completion of basic education, by district type (dzud 1999–2002)
Table 11 Placebo tests of the impact of the dzuds on completion of mandatory education, OLS
Table 12 The impact of the 1999–2002 dzuds on completion of mandatory education (full set of controls shown), OLS
Table 13 The impact of the 2009/2010 dzud on completion of mandatory education (full set of controls shown), OLS

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Groppo, V., Kraehnert, K. The impact of extreme weather events on education. J Popul Econ 30, 433–472 (2017). https://doi.org/10.1007/s00148-016-0628-6

Download citation

Keywords

  • Children
  • Education
  • Extreme weather events
  • Mongolia

JEL Classification

  • I25
  • Q54
  • O12