A cyclone passes, buildings are reconstructed, children return to school, and yet something doesn’t always come back. A study by the Asian Development Bank covering seven Asian countries establishes that exposure to disasters during the first 1,000 days from conception is associated with lasting effects on school enrollment and mathematics skills. However, it does not measure the effect of physical school reconstruction. The most striking result concerns floods, which rank among the categories producing the most severe effects on school enrollment and mathematics results—two important components of human capital strongly associated with subsequent educational, professional, and economic trajectories.

Key Points

  • According to the Asian Development Bank, the most robust negative effects concern disasters experienced during the first 1,000 days from conception, rather than throughout the entire school-age period.
  • Among the categories analyzed, floods present the most robust negative effects on school enrollment and mathematics scores. Severe floods have the largest estimated early effect on enrollment, but not a statistically established average effect that is the most negative on mathematics.
  • The loss extends over time: the study covers seven Asian countries and documents persistent effects from certain early exposures, without measuring reconstruction periods.
  • The timing of exposure plays a key role: associations are stronger for exposures during the first 1,000 days from conception than for recent or mid-childhood exposures.
  • School protection constitutes a resilience policy distinct from physical reconstruction, engaging the responsibility of governments and donors.

One Flood Erases Two Years of Mathematics

Natural disasters have always had measurable economic effects. We count destroyed homes, collapsed bridges, lost harvests. What we measure less well is what disappears in the minds of children growing up in affected areas. The Asian Development Bank study published in February 2025 provides a more precise answer than what the literature had produced thus far, by cross-referencing disaster data with measures of school enrollment and cognitive skills in seven countries: Bangladesh, Kyrgyz Republic, Mongolia, Nepal, Pakistan, Thailand, and Turkmenistan.

The clearest result concerns floods. Among the categories actually analyzed—any disaster, floods, severe disasters, and severe floods—floods present the most robust negative effects on school enrollment rates and mathematics competencies. Severe floods show the largest estimated early effect on enrollment, but their average effect on mathematics is not statistically established as the most negative. The study does not separately compare cyclones, droughts, and earthquakes. But the frequency of floods and their ability to disrupt local infrastructure for weeks—cut-off roads, schools turned into shelters, displaced teachers—can create interruptions at the moment when learning is being built.

Fundamental learning has the particular characteristic of being cumulative. Understanding fractions requires mastery of multiplication. Mastery of multiplication requires having automated addition. A child deprived of part of their schooling early in their education can encounter additional difficulties in subsequent years, because teachers advance on the basis of what students are supposed to know.

Timing Matters More Than Intensity

One of the most important contributions of the study is to distinguish not only the categories of disasters, but the moment at which they occur in the life trajectory. It compares exposure during the first 1,000 days from conception, mid-childhood, and recent years. The effects associated with very early exposures are more marked than those observed for mid-childhood or recent exposures.

This finding is consistent with what research in education economics has documented since James Heckman’s work on early investment: cognitive skills develop within a time window where inputs have higher returns than later. Heckman demonstrated this for preschool education policies; the ADB study shows, for negative shocks, that exposure during the first 1,000 days from conception generally has more negative effects than recent or mid-childhood exposure.

This result has direct implications for reconstruction policies. Building a stronger school is necessary, but resuming classes does not guarantee by itself the recovery of interrupted learning. Physical reconstruction and learning support do not necessarily follow the same timeline.

Seven Countries, a Regional Phenomenon

The scope of the study deserves attention. Bangladesh, the Kyrgyz Republic, Mongolia, Nepal, Pakistan, Thailand, and Turkmenistan do not form a homogeneous group. Their education systems, development levels, exposure to natural disasters, and institutional capacities differ considerably. The study thus allows observation of the effects of certain early exposures in varied contexts, without necessarily supporting a comparison with the Philippines, which does not appear in the sample.

South and Southeast Asia is one of the world’s regions most exposed to natural disasters. This exposure will intensify with climate change: IPCC projections indicate an intensification of extreme precipitation in several Asian regions, which can translate into more frequent and more severe floods. The region accounts for some of the world’s largest school populations. The issues of social mobility at stake are massive, and as other work shows, spending more per student is not enough to reduce school inequalities if the structural causes of disruption are not addressed.

The ADB published, in February 2025, a work devoted to strengthening the resilience of education systems. This work calls for better preparing school systems for disruptions, particularly climate-related ones, without constituting a verified assessment of the institution’s past financing nor a precise institutional recognition of insufficient integration of climate resilience in its previous investments.

Poor Children Pay the Price Twice

The effects of disasters are not necessarily distributed uniformly. In affected areas, children from the most vulnerable households may be particularly exposed to school interruptions. Several mechanisms add together.

First, poor families often live in the most exposed zones—floodplains, unprotected coastlines, rural areas far from infrastructure. Physical exposure to risk is itself unevenly distributed by income. Second, when disaster strikes, poor families have fewer resources to compensate for the school interruption: no tutoring, no parents able to ensure pedagogical continuity at home, sometimes school withdrawal to send children to work during the reconstruction phase. Third, school facilities in poor areas are often the least well built and equipped to reopen quickly.

The result can be a double penalty: more exposed to the shock, some children are also less capable of recovering from it. A child born in a flood-prone area in Bangladesh can thus face a greater risk of experiencing interrupted schooling. This mechanism can transform a temporary shock into long-term inequality, what education economists call a scar.

Protecting Schools During Disaster, Not Just After

Facing this diagnosis, several countries and organizations have begun thinking about approaches that do not merely rebuild. The issue is to maintain pedagogical continuity during the disaster itself, and to organize targeted remediation in the weeks and months that follow.

The idea of protecting cognitive infrastructure aligns with a broader debate on the nature of investments in climate adaptation. Building dikes, raising buildings, installing early warning systems—all this protects lives and property. Protecting educational trajectories is protecting a form of human capital whose destruction is not immediately visible but whose effects can be measured over decades of income, mobility, and individual capabilities.

The Coming Decades Will Make This More Urgent

Climate change will increase the frequency and intensity of floods in several Asian regions. An increase in climate hazards can increase the number of children exposed and intensify pressure on education systems. By itself, it does not allow the conclusion that the automatic emergence of a deficit cohort every decade is inevitable, nor that there is a complete absence of catch-up.

Two trajectories are emerging, conditional on policies adopted. In the first, countries and donors continue to treat education as infrastructure to be reconstructed after disaster, without protecting pedagogical continuity during and immediately after the shock. In this scenario, school interruptions could weigh more heavily on the trajectories of exposed children and reinforce gaps between territories—a phenomenon we already see at work in other technological domains, where AI amplifies the inequalities it promised to erase for lack of deliberate access policies.

In the second trajectory, governments and institutions like the ADB integrate the resilience of education systems as a full dimension of their climate adaptation strategies. This requires at minimum three types of measures: emergency education protocols allowing continuity during the shock; targeted remediation programs in the months following the disaster; and investments in teacher training to manage disruptions pedagogically.

What would allow us to distinguish between the two trajectories before 2030 are a few observable signals. If new programs integrate learning indicators after disasters, not just physical reconstruction indicators, that would be a sign that the diagnosis is beginning to translate into policy. Similarly, the number of countries that adopt tested and evaluated emergency education plans constitutes an indicator of how quickly the education sector is adapting to a climate reality that does not wait.

Children entering primary school in 2025 in Bangladesh’s deltas, in South Asian floodplains, or in exposed coastal zones will likely be adults when floods are more frequent than today. The skills they will have will depend in part on the intensity of the next shocks, but also on the educational protections their governments choose, or do not choose, to put in place.

Sources

  1. Asian Development Bank, Are Natural Disasters Disastrous for Education? Evidence from Seven Asian Countries, February 2025, https://www.adb.org/publications/are-natural-disasters-disastrous-for-education-evidence-from-seven-asian-countries
  2. Asian Development Bank, An Evaluation of ADB’s Support for Education in Asia and the Pacific, March 2025, no verified URL
  3. James Heckman, work on returns to early educational investment, University of Chicago, publications accessible via heckmanequation.org
  4. Asian Development Bank and ADB Institute, Are Natural Disasters Disastrous for Education? Evidence from Seven Asian Countries, full text of Working Paper 1492, https://yujiezhangecon.github.io/files/DisasterEducationAsiaMicsEmDat_ZhangEtal.pdf
  5. IPCC, AR6, Working Group II, Chapter 10, Asia, https://www.ipcc.ch/report/ar6/wg2/chapter/chapter-10/
  6. IPCC, AR6, Working Group I, Chapter 11, Weather and Climate Extreme Events in a Changing Climate, https://www.ipcc.ch/report/ar6/wg1/chapter/chapter-11/
  7. Asian Development Bank, Building Resilience in Education Systems, February 2025, https://www.adb.org/publications/building-resilience-in-education-systems
  8. World Bank, Shock Waves: Managing the Impacts of Climate Change on Poverty, https://documents1.worldbank.org/curated/en/260011486755946625/pdf/ShockWaves-FullReport.pdf
  9. IPCC, AR6, Working Group I, Chapter 12, Climate Change Information for Regional Impact and for Risk Assessment, https://www.ipcc.ch/report/ar6/wg1/chapter/chapter-12/
  10. IPCC, AR6, Working Group I, Technical Summary, https://www.ipcc.ch/report/ar6/wg1/chapter/technical-summary/
  11. World Bank, What Is Learning Poverty?, https://www.worldbank.org/en/topic/education/brief/what-is-learning-poverty