School Closures and Reopening in the Pandemic

School clo­sures due to Covid-19 have left 1.6 bil­lion stu­dents with­out in-per­son class­es for a pro­longed pe­ri­od. The lion’s share of those stu­dents is in de­vel­op­ing coun­tries, with lim­it­ed ac­cess to the de­vices and con­nec­tiv­i­ty nec­es­sary to fa­cil­i­tate re­mote learn­ing. In de­cid­ing whether to al­low in-per­son class­es to re­sume, pol­i­cy­mak­ers face a trade off be­tween the po­ten­tial health costs as­so­ci­at­ed with the of re­open­ing schools dur­ing the pan­dem­ic and the po­ten­tial ed­u­ca­tion­al costs of keep­ing them shut for longer still. How­ev­er, ev­i­dence on the mag­ni­tude of these costs in de­vel­op­ing coun­tries is lim­it­ed. Lack­ing suf­fi­cient data to in­form gov­ern­men­tal de­ci­sions, the pub­lic de­bate has cen­tered on whether to re­open schools or not, rather than on how to re­open them safe­ly, putting the fu­ture of chil­dren and ado­les­cents at risk.

The Sus­tain­able De­vel­op­ment Goals of this Pro­ject
good health and well-beeingquality education

First­ly, we doc­u­ment first-hand the ed­u­ca­tion­al im­pacts of school clo­sures and re­open­ing in a de­vel­op­ing coun­try dur­ing the pan­dem­ic. Tak­ing ad­van­tage of the fact that São Paulo fea­tured in-per­son class­es for most of the first school quar­ter of 2020, but not af­ter that, we find that dropout risk in­creased by 365% dur­ing school clo­sures in 2020. While risk in­creased with lo­cal dis­ease ac­tiv­i­ty, most of it can be at­trib­uted di­rect­ly to the ab­sence of in-per­son class­es: we es­ti­mate that dropout risk in­creased by no less than 247% across the State, even at the low end of the dis­tri­bu­tion of per capi­ta COVID-19 cas­es. Stu­dents only achieved 27.5% of the learn­ing out­comes of the typ­i­cal year equiv­a­lent while schools were closed, with learn­ing loss­es con­cen­trat­ed on vul­ner­a­ble stu­dents (girls, black and brown stu­dents) and on vul­ner­a­ble schools (those lo­cat­ed in the poor­est neigh­bor­hoods, and that did not of­fer on­line aca­d­e­m­ic ac­tiv­i­ties be­fore the pan­dem­ic).

Au­tho­riz­ing schools to par­tial­ly re­open for in-per­son class­es at the 4th quar­ter of 2020 in­creased high-school stu­dents’ test scores by 20%, bring­ing learn­ing loss­es to rough­ly 2/3 (com­pared to 3/4 in mu­nic­i­pal­i­ties that did not au­tho­rize schools to re­open). Hav­ing said that, re­open­ing schools for this lim­it­ed pe­ri­od could not mit­i­gate the dra­mat­ic in­crease in dropout risk dur­ing school clo­sures. More­over, the re­open­ing of schools only for op­tion­al ac­tiv­i­ties was in­suf­fi­cient to mit­i­gate learn­ing loss­es among mid­dle-school stu­dents.

Sec­ond­ly, we doc­u­ment the po­ten­tial con­tri­bu­tion of school re­open­ing dur­ing that pe­ri­od on lo­cal dis­ease ac­tiv­i­ty. We find that school re­open­ing did not in­crease Covid-19 in­ci­dence or mor­tal­i­ty on av­er­age, up to 12 weeks af­ter re­open­ing, even in mu­nic­i­pal­i­ties with low-qual­i­ty school in­fra­struc­ture, low per capi­ta in­come, high se­nior pop­u­la­tion share, or those most se­vere­ly af­fect­ed by the pan­dem­ic. School re­open­ing in the pan­dem­ic did not in­crease mu­nic­i­pal-lev­el Covid-19 in­ci­dence or deaths not only be­cause school com­mu­ni­ties rep­re­sent a small frac­tion of the over­all pop­u­la­tion, but also be­cause coun­ter­fac­tu­al mo­bil­i­ty dur­ing the pan­dem­ic was al­ready sub­stan­tial: Google data at­tests that, by the time schools re­opened for in-per­son ac­tiv­i­ties, lo­cal mo­bil­i­ty was sim­i­lar to that be­fore the pan­dem­ic – even in mu­nic­i­pal­i­ties that did not au­tho­rize schools to re­open in 2020 – ren­der­ing the mar­gin­al health ben­e­fits of keep­ing schools closed neg­li­gi­ble.

The loop child and ado­les­cent un­der-ed­u­ca­tion, as a re­sult of in­creased dropout risk and learn­ing loss­es with­out in-per­son class­es.

Break­ing the loop: al­low­ing in-per­son class­es to re­turn un­der safe school re­open­ing pro­to­cols to en­sure that every child and youth have ac­cess to ad­e­quate ed­u­ca­tion

School Clo­sures and Re­open­ing in the Pan­dem­ic

The study finds that in-per­son ac­tiv­i­ties did not con­tribute to ag­gre­gate Covid-19 cas­es or deaths be­tween Oc­to­ber and De­cem­ber 2020 in São Paulo State, Brazil – large­ly be­cause mo­bil­i­ty was al­ready very close to pre-pan­dem­ic lev­els in the State.

  • Sta­tusOn­go­ing
  • Coun­tryBrazil
  • Pro­gram areaNo Pover­ty, Health Well­be­ing, Ed­u­ca­tion, COVID-19
  • Top­icsCOVID-19, School clo­sure, learn­ing costs, health costs
  • Part­nersSec­re­taria de Ed­u­cação do Es­ta­do de São Paulo, MOV­VA
  • Time­line2020-2021
  • Study TypeRan­dom­ized eval­u­a­tion and qua­si-ex­per­i­men­tal de­sign
  • Sam­ple Size~200,000 stu­dents in the ex­per­i­men­tal in­ter­ven­tion and 645 mu­nic­i­pal­i­ties in the qua­si-ex­per­i­men­tal de­sign

Re­search Team

Carlos Belchior


João Cossi