Teacher transfers and the disruption of teacher staffing in the city of Sao Paulo

Elacqua, G ego y; Rosa, Leona do Wo king Pape Teache ans e s and he dis up ion o eache s a ing in he ci y o Sao Paulo IDB Wo king Pape Se ies, No. IDB-WP-01437 P o ided in Coope a ion wi h: In e -Ame ican De elopmen Bank (IDB), Washing on, DC Sugges ed Ci a ion: Elacqua, G ego y; Rosa, Leona do (2023) : Teache ans e s and he dis up ion o eache s a ing in he ci y o Sao Paulo, IDB Wo king Pape Se ies, No. IDB-WP-01437, In e - Ame ican De elopmen Bank (IDB), Washing on, DC, h ps://doi.o g/10.18235/0004737 This Ve sion is a ailable a : h ps://hdl.handle.ne /10419/289964 S anda d-Nu zungsbedingungen: Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen Zwecken und zum P i a geb auch gespeiche und kopie we den. Sie dü en die Dokumen e nich ü ö en liche ode komme zielle Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich machen, e eiben ode ande wei ig nu zen. So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen (insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en, gel en abweichend on diesen Nu zungsbedingungen die in de do genann en Lizenz gewäh en Nu zungs ech e. Te ms o use: Documen s in EconS o may be sa ed and copied o you pe sonal and schola ly pu poses. You a e no o copy documen s o public o comme cial pu poses, o exhibi he documen s publicly, o make hem publicly a ailable on he in e ne , o o dis ibu e o o he wise use he documen s in public. I he documen s ha e been made a ailable unde an Open Con en Licence (especially C ea i e Commons Licences), you may exe cise u he usage igh s as speci ied in he indica ed licence. h ps://c ea i ecommons.o g/licenses/by-nc-nd/3.0/igo/legalcode Teache ans e s and he dis up ion o eache s a ing in he Ci y o Sao Paulo G ego y Elacqua Leona do Rosa IDB WORKING PAPER SERIES Nº IDB-WP-01437 Feb ua y 2023 Social Sec o Depa men Educa ion Di ision In e -Ame ican De elopmen Bank Feb ua y 2023 Teache ans e s and he dis up ion o eache s a ing in he Ci y o Sao Paulo G ego y Elacqua Leona do Rosa Ca aloging-in-Publica ion da a p o ided by he In e -Ame ican De elopmen Bank Felipe He e a Lib a y Elacqua, G ego y M., 1972- Teache ans e s and he dis up ion o eache s a ing in he ci y o Sao Paulo / G ego y Elacqua, Leona do Rosa. p. cm. — (IDB Wo king Pape Se ies ; 1437) Includes bibliog aphic e e ences. 1. Teache s-Sala ies, e c.-B azil. 2. Teache u no e -B azil. 3. Teache s-Selec ion and appoin men -B azil. I. Rosa, Leona do. II. In e -Ame ican De elopmen Bank. Educa ion Di ision. III. Ti le. IV. Se ies. IDB-WP-1437 JEL Codes: I21, I24, I25, I28 Keywo ds: eache s, eache assignmen , eache ans e s, cen alized assignmen , B azil Copy igh © In e -Ame ican De elopmen Bank. This wo k is licensed unde a C ea i e Commons IGO 3.0 A ibu ion- NonComme cial-NoDe i a i es (CC-IGO BY-NC-ND 3.0 IGO) license (h p://c ea i ecommons.o g/licenses/by-nc-nd/3.0/igo/ legalcode) and may be ep oduced wi h a ibu ion o he IDB and o any non-comme cial pu pose, as p o ided below. No de i a i e wo k is allowed. Any dispu e ela ed o he use o he wo ks o he IDB ha canno be se led amicably shall be submi ed o a bi a ion pu suan o he UNCITRAL ules. The use o he IDB's name o any pu pose o he han o a ibu ion, and he use o IDB's logo shall be subjec o a sepa a e w i en license ag eemen be ween he IDB and he use and is no au ho ized as pa o his CC-IGO license. Following a pee e iew p ocess, and wi h p e ious w i en consen by he In e -Ame ican De elopmen Bank (IDB), a e ised e sion o his wo k may also be ep oduced in any academic jou nal, including hose indexed by he Ame ican Economic Associa ion's EconLi , p o ided ha he IDB is c edi ed and ha he au ho (s) ecei e no income om he publica ion. The e o e, he es ic ion o ecei e income om such publica ion shall only ex end o he publica ion's au ho (s). Wi h ega d o such es ic ion, in case o any inconsis ency be ween he C ea i e Commons IGO 3.0 A ibu ion-NonComme cial-NoDe i a i es license and hese s a emen s, he la e shall p e ail. No e ha link p o ided abo e includes addi ional e ms and condi ions o he license. The opinions exp essed in his publica ion a e hose o he au ho s and do no necessa ily e lec he iews o he In e -Ame ican De elopmen Bank, i s Boa d o Di ec o s, o he coun ies hey ep esen . h p://www.iadb.o g 2023 Teache ans e s and he dis up ion o eache s affing in he Ci y o Sao Paulo G ego y Elacqua, IADB∗ Leona do Rosa, IEPS† Feb ua y 10, 2023 Abs ac This pape analyzes p e e ences o ce ain school a ibu es among in-se ice eache s. We explo e a cen alized ma ching p ocess in he ci y o Sao Paulo ha eache s mus use when ans e ing schools. Because eache s ha e o lis and ank hei p e e ences o schools, we can es ima e he desi abili y o school a ibu es using a ank-o de ed logi model. We show ha he school’s dis ance om he eache ’s home, i s a e age es sco es, and eache composi ion play a cen al ole in eache p e e ences. Fu he mo e, we documen ha p e e ences a y acco ding o eache cha ac e is ics such as gende , ace, age, and academic subjec . ∗In e -Ame ican De elopmen Bank, [email p o ec ed] †Ins i u o de Es udos pa a Pol´ı icas de Sa´ude,[email p o ec ed] 1 1 In oduc ion Teache s a e he mos c ucial inpu in he educa ion p oduc ion unc ion (Che y e al., 2014) and hei impac is g ea es among he mos disad an aged s uden s (A aujo e al., 2016). Howe e , a ac ing and e aining qualified eache s o low-pe o ming schools ha se e low-income and mino i y s uden s is one o he bigges challenges school sys ems ace. Teache s o en so acco ding o s uden socioeconomic le el (Lank- o d e al., 2002), school acial composi ion (Hanushek e al., 2004), geog aphic loca ion (Be oni e al., 2022), and pupil achie emen (K ieg e al., 2016). They also p e e es- ablishmen s ha ha e be e wo king condi ions (Boyd e al., 2005) and a e close o hei homes o o whe e hey wen o college (Be oni e al., 2022; Boyd e al., 2005). As a esul o eache so ing, disad an aged and low-pe o ming s uden s a e less likely o a end classes wi h qualified eache s (Jackson and B uegmann, 2009), which has a nega i e impac on hei lea ning (A aujo e al., 2016). Wi h ew excep ions, esea ch in his a ea ocuses on how eache s so acco ding o specific school cha ac e is ics. While some s udies use su ey e idence o shed ligh on his opic, o he s explo e he impac o eache choices on so ing. The scan e idence on eache s’ p e e ences in La in Ame ica is la gely ela ed o he cha ac e is ics ha d i e he decisions o new eache candida es (Be oni e al., 2022; Rosa, 2019). Ve y li le wo k has been done (G issom e al., 2014 being a no able excep ion) on how eache so ing is affec ed by he p e e ences o in-se ice eache s, who make up he majo i y o eache s in he labo ma ke . In his pape , we analyze he p e e ences o in-se ice eache s in he ci y o Sao Paulo. Using a ich adminis a i e da ase ha p o ides unique in o ma ion on in- se ice eache s’ exi decisions and o de ed ans e p e e ences, we aim o answe he ollowing ques ions: (i) wha school ac o s a e associa ed wi h eache s’ desi e o ans- e away?; (ii) in sea ching o a new school, wha cha ac e is ics do eache s look o ? Using a ank-o de logi model, we examine how in-se ice eache ans e applican s e alua e diffe en school cha ac e is ics in he es ablishmen s hey lea e and when ank- ing hei p e e ed ans e s. Es ima ing wo ke p e e ences is challenging o any occupa ion, and eache p e - e ences a e no diffe en . In a compe i i e job ma ke , eache s and school p e e ences in e ac , making i difficul o disen angle p e e ences when obse ing ma ches be ween 2 eache s and schools. Analyzing his wo-sided ma ching p ocess equi es some assump- ions, which can limi unde s anding. In a s anda d case, we obse e he choices made by eache s, bu we do no obse e eache choice se s. In o he wo ds, we do no know he op ions eache s ha e a ailable o hem and he ac o s hey a e conside ing when choosing a school, which makes i complica ed o es ima e eache p e e ences. Fu he mo e, when conside ing he eache labo ma ke wi hin a single educa ion sys em, school cha ac e is ics end o be homogeneous. The lack o a ia ion in school ea u es limi s ou abili y o d aw conclusions since we canno es ima e eache p e e ences o many school ac o s (e.g., eache wages). Finally, eache p e e ences migh diffe be ween no ice and in-se ice eache s. While no ice eache s likely ha e conside ably less in o ma ion abou school cha ac e is ics because hey a e no pa o he wo k o ce, in-se ice eache s ha e mo e expe ience in schools, p esumably ha e mo e in o ma ion abou he local schooling ma ke , and compa e hei cu en schools wi h o he po en ial wo kplaces. Ou case s udy offe s an in e es ing oppo uni y o add ess some o hese challenges. While Sao Paulo is he la ges ci y in B azil – and one o he la ges in he wo ld – i has a highly cen alized municipal school sys em. The Depa men o Educa ion manages he ans e p ocess ac oss es ablishmen s, which we use o analyze eache p e e ences. Once a yea , eache s can eques a ans e om hei cu en school o ano he school in he municipal sys em. The Depa men o Educa ion egis e s he eache s’ eques s and coun s acancies gene a ed by a i ion e en s (e.g., e i emen o eache lea es). The go e nmen publishes openings on an online pla o m and asks he eache s who eques ed a ans e o ank hei school p e e ences. They can ank an unlimi ed numbe o acancies. Finally, eache s a e ma ched wi h hei p e e ed schools using a de e ed accep ance algo i hm. I wo eache s apply o he same school, he mechanism ma ches he school o he eache wi h mo e poin s (based on a combi- na ion o expe ience and educa ion le el), wi h he ma ching p ocess con inuing o e a se ies o ounds. A guably, his cen alized p ocess mi iga es mos conce ns ela ed o es ima ing eache p e e ences. Fi s , we do no need o disen angle eache p e e ences and school p e e ences. The go e nmen s anda dizes school (o p incipal) p e e ences in he ma ching p ocess. Second, because he en i e ma ching p ocess occu s on a cen alized pla o m, we know he op ions a ailable o eache s when hey ank hei p e e ences. 3 Thus, we can obse e hei choice se s in a compe i i e ma ke . Thi d, he school sys- em and he ci y a e e y la ge, wi h mo e han fi e hund ed schools. The e o e, we obse e conside able a ia ion ac oss school cha ac e is ics, including quali y, a e age socioeconomic le el, loca ion, and mone a y incen i es. Finally, we conside in-se ice eache s and no no ice eache s, meaning ha he eache s in ou sample a e mo e in o med and capable o obse ing and compa ing school cha ac e is ics. When analyzing in-se ice eache s asking o ans e s, ou esul s indica e ha eache s a e mo e likely o ans e ou o schools wi h lowe es sco es, a highe pe cen age o low-SES s uden s, and ha a e a he away om hei homes. We use a ank-o de ed logi model o es ima e eache p e e ences o new schools. Ou es ima es show ha dis ance om home o school is he s onges p edic o o eache p e e ences. Howe e , o he school cha ac e is ics a e impo an as well: a school’s a e age es sco es a e a e y s ong p edic o o eache p e e ences. School o ganiza ion is ano he impo an ac o behind eache p e e ences, as eache s a e mo e likely o a oid schools wi h many no ice eache s and ewe class ooms. In e es - ingly, a e con olling o dis ance and es sco es, he a e age socioeconomic le el o he s uden body is no ela ed o eache p e e ences. Ou s udy in e sec s wi h he g owing li e a u e on eache p e e ences, eache u no e , and an eme ging body o wo k ha explo es eache so ing in de eloping coun ies. In he eache p e e ence li e a u e, mos s udies analyze cases whe e he eache labo ma ke migh be classified as a wo-sided ma ching (Bonhomme e al., 2016; Boyd e al., 2013; Engel e al., 2014), and he choice se o eache p e e ence anking migh no be e y well-defined. One s udy mi iga es his conce n by analyzing a disc e e-choice expe imen wi h eache s, con olling he choice se en i onmen and in oducing a ia ion in he school and job-con ac ea u es (Johns on, 2020). He e, we p esen a diffe en case and add ess key challenges in he s udy o eache p e e ences, as discussed abo e. Ou pape also con ibu es o unde s anding eache u no e .1In his a ea, he mos closely ela ed s udy o ou s is Boyd e al. (2011). The au ho s examine he eache ans e sys em in New Yo k Ci y, which is cen alized in a simila manne o Sao Paulo. We build on his esea ch by explo ing diffe en school cha ac e is ics and 1Allen e al. (2018), C. Clo el e e al. (2008), C. T. Clo el e e al. (2008), and C. T. Clo el e e al. (2004), Falch and S øm (2005). 4 a diffe en con ex . In con as o Boyd e al. (2011) bu consis en wi h o he findings in he li e a u e (Bonhomme e al., 2016), we show ha a e age school es sco es a e impo an p edic o s o eache ans e s. Finally, we build on he g owing li e a u e in de eloping coun ies ha explo es he ac o s ha influence eache so ing and he mos effec i e policies o a ac ing and e aining eache s in ha d- o-s aff schools (Be oni e al., 2022; E ans, Acos a, e al., 2021). P e ious wo k ocuses on he influence o new eache p e e ences in so ing. Fo ins ance, Be oni e al. (2022) analyze eache p e e ences in he cen alized eache selec ion and assignmen sys em in Pe u and show ha mo e expe ienced and highe - pe o ming eache s p e e schools in mo e ad an aged u ban a eas, loca ed in places close o whe e hey a ended college. Rosa (2019) s udies he ec ui men o eache s in Sao Paulo and documen s ha loca ion and es sco es a e impo an de e minan s o ini ial eache decisions. Unlike p e ious s udies ha examine he decisions o no ice eache s, we look a in-se ice eache s and hei decisions o ans e o a diffe en school. 2 Teache ans e s in he ci y o Sao Paulo App oxima ely 50,000 eache s wo k in p ima y and middle schools in he ci y o Sao Paulo. The Depa men o Educa ion hi es all eache s; schools do no pa icipa e in he hi ing p ocess. To be hi ed as an in-se ice eache , candida es mus pass a cen alized assessmen . Once hey pass he exam, he depa men in i es eache s o choose a school om a lis o acancies ha hey publish online. Teache s can decide o lea e he school sys em a any ime. A he end o each school yea , in-se ice eache s can eques a ans e o ano he school wi hin he school sys em using he ci y’s cen alized sys em. In his ans e p ocess, in-se ice eache s ha e p io i y o e no ice eache s as hey a e allowed o choose fi s . Ou analysis ocuses on hese ans e p ocesses. Specifically, each yea , manage s in he Depa men o Educa ion consolida e he numbe o eache s ha exi he school sys em and he numbe o eache s eques ing a ans e o compu e he o al numbe o acancies pe school. Then, hey open he concu so de emo¸c˜ao, a p ocess ha allows eache s o eques ans e s ac oss public municipal schools. Teache s who eques a ans e mus ank, using an online pla o m, hei school p e e ences based on he official lis o acancies published by 5 We find ha , when ans e ing, eache s a e mo e likely o lea e schools wi h lowe academic pe o mance, a highe p opo ion o low-SES s uden s, and a g ea e numbe o no ice eache s, as well as es ablishmen s loca ed in a a ela and a he away om eache s’ homes. In con as , he schools ha eache s anked as hei op choice in ans e eques s end o be la ge ones wi h highe es sco es, a lowe p opo ion o low-SES s uden s, and ewe no ice eache s, as well as no loca ed in a a ela and close o hei homes. I ul ima ely assigned (28% we e no g an ed a ans e , mainly due o ma ke conges ion), he school ends o be somewhe e in he middle be ween hei op p e e ence and hei cu en school in e ms o he cha ac e is ics lis ed abo e. We also find ha p e e ences a y ac oss diffe en g oups o in-se ice eache s: loca ion is pa icula ly impo an o emale and olde eache s (>30 yea s). Women a e also mo e likely o choose schools wi h a lowe pe cen age o disad an aged s uden s. Black eache s a e less likely o choose o ans e o schools loca ed in a a ela and mo e likely o p e e schools ha offe mone a y incen i es. In addi ion, we find ha younge eache s (<30 yea s o age) a e mo e likely o eques ans e s o schools wi h mo e expe ienced eache s. Mo e a ac i e wages a ha d- o-s aff schools, which a e mo e likely o be low- pe o ming and se e mo e disad an aged s uden s, may influence eache ans e de- cisions. To add ess his issue, he Ci y o Sao Paulo has in oduced mone a y incen i es o wo k in specific neighbo hoods. Howe e , he wage bonus ep esen s a small pe cen - age o a eache ’s sala y and is hus unlikely o be sufficien o o e come eache s’ dislike o he cha ac e is ics o schools loca ed in disad an aged neighbo hoods. Indeed, only Black eache s, who may be mo e likely o eside in o close o neighbo hoods wi h wage bonuses, a e mo e likely o eques ans e s o schools wi h mone a y incen i es. The Ci y o Sao Paulo may benefi om e iewing he cu en mone a y incen i e p o- g am. Fu he s udies could e alua e he effec o he wage bonus mo e di ec ly, as well as i s impac on eache e en ion and u no e and on he eligible schools and neighbo hoods. Go e nmen s in de eloping coun ies ha e adop ed a ious s a egies o inc ease he supply o eache s in ha d- o-s aff schools, including mone a y incen i es, as e acks o p omo ion, and manda o y o a ions (E ans, Acos a, e al., 2021). In ligh o he s ong p e e ence o es ablishmen s close o home, some school sys ems offe ans- po a ion and housing subsidies (Puga ch & Sch oede , 2014). Mo e ecen ly, go e n- 12 men s in La in Ame ica ha e been expe imen ing wi h beha io al nudges o a ac eache s o disad an aged schools (Ajzenman e al., 2020; Ajzenman e al., 2021). The esea ch shows ha simple low-cos in e en ions a e effec i e a encou aging eache s o apply o ha d- o-s aff schools, offe ing a p omising al e na i e o financial incen i es a a ime when economic g ow h is low and educa ion budge s a e igh . While his app oach has shown some success in ge ing no ice eache s o apply o ha d- o-s aff schools, u u e wo k migh conside es ing simila s a egies o mi iga e he so ing o mo e expe ienced eache s seeking ans e s. 13 Re e ences Ajzenman, N., Be oni, E., Elacqua, G., Ma o a, L., & M´endez Va gas, C. (2020). Al uism o money? educing eache so ing using beha io al s a egies in pe u. Reducing Teache So ing using Beha io al S a egies in Pe u (July 30, 2020). Ajzenman, N., Elacqua, G., Ma o a, L., & Olsen, A. (2021). O de effec s and employ- men decisions: Expe imen al e idence om a na ionwide p og am. Allen, R., Bu gess, S., & Mayo, J. (2018). The eache labou ma ke , eache u no e and disad an aged schools: New e idence o england. Educa ion Economics, 26(1), 4–23. A aujo, M. C., Ca nei o, P., C uz-Aguayo, Y., & Schady, N. (2016). Teache quali y and lea ning ou comes in kinde ga en. The Qua e ly Jou nal o Economics, 131(3), 1415–1453. Be oni, E., Elacqua, G., Hincapi´e, D., M´endez, C., & Pa edese, D. (2022). Teache s’ p e e ences o p oximi y and he implica ions o s affing schools: E idence om pe u. Educa ion Finance and Policy, 1–32. Bonhomme, S., Joli e , G., & Leu en, E. (2016). School cha ac e is ics and eache u no e : Assessing he ole o p e e ences and oppo uni ies. The Economic Jou nal,126(594), 1342–1371. Boyd, D., Lank o d, H., Loeb, S., Ron eld , M., & Wyckoff, J. (2011). The ole o eache quali y in e en ion and hi ing: Using applica ions o ans e o unco e p e e ences o eache s and schools. Jou nal o Policy Analysis and Managemen , 30(1), 88–110. Boyd, D., Lank o d, H., Loeb, S., & Wyckoff, J. (2005). Explaining he sho ca ee s o high-achie ing eache s in schools wi h low-pe o ming s uden s. Ame ican economic e iew,95(2), 166–171. Boyd, D., Lank o d, H., Loeb, S., & Wyckoff, J. (2013). Analyzing he de e minan s o he ma ching o public school eache s o jobs: Disen angling he p e e ences o eache s and employe s. Jou nal o Labo Economics,31(1), 83–117. Che y, R., F iedman, J. N., & Rockoff, J. E. (2014). Measu ing he impac s o eache s ii: Teache alue-added and s uden ou comes in adul hood. Ame ican economic e iew,104(9), 2633–79. Clo el e , C., Glennie, E., Ladd, H., & Vigdo , J. (2008). Would highe sala ies keep eache s in high-po e y schools? e idence om a policy in e en ion in no h ca olina. Jou nal o Public Economics,92(5-6), 1352–1370. 14 Clo el e , C. T., Glennie, E. J., Ladd, H. F., & Vigdo , J. L. (2008). Teache bonuses and eache e en ion in low-pe o ming schools: E idence om he no h ca - olina 1,800 eache bonusp og am.Public Finance Re iew,36(1), 63–87. Clo el e , C. T., Ladd, H. F., Vigdo , J. L., & Diaz, R. A. (2004). Do school ac- coun abili y sys ems make i mo e difficul o low-pe o ming schools o a ac and e ain high-quali y eache s? Jou nal o Policy Analysis and Managemen , 23(2), 251–271. Elacqua, G., & Ma o a, L. (2020). Is wo king one job be e han many? assessing he impac o mul iple school jobs on eache pe o mance in io de janei o. Economics o Educa ion Re iew,78, 102015. Engel, M., Jacob, B. A., & Cu an, F. C. (2014). New e idence on eache labo supply. Ame ican Educa ional Resea ch Jou nal,51(1), 36–72. E ans, D. K., Acos a, A. M., e al. (2021). How o ec ui eache s o ha d- o-s aff schools: A sys ema ic e iew o e idence om low-and middle-income coun ies. Cen e o Global De elopmen . Falch, T., & S øm, B. (2005). Teache u no e and non-pecunia y ac o s. Economics o Educa ion Re iew,24(6), 611–631. G issom, J. A., Loeb, S., & Nakashima, N. A. (2014). S a egic in olun a y eache ans e s and eache pe o mance: Examining equi y and efficiency. Jou nal o Policy Analysis and Managemen ,33(1), 112–140. Hanushek, E. A., Kain, J. F., & Ri kin, S. G. (2004). Why public schools lose eache s. Jou nal o human esou ces,39(2), 326–354. Jackson, C. K., & B uegmann, E. (2009). Teaching s uden s and eaching each o he : The impo ance o pee lea ning o eache s. Ame ican Economic Jou nal: Ap- plied Economics,1(4), 85–108. Johns on, A. C. (2020). Teache p e e ences, wo king condi ions, and compensa ion s uc u e. K ieg, J. M., Theobald, R., & Goldhabe , D. (2016). A oo in he doo : Explo ing he ole o s uden eaching assignmen s in eache s’ ini ial job placemen s. Educa- ional E alua ion and Policy Analysis,38(2), 364–388. Lank o d, H., Loeb, S., & Wyckoff, J. (2002). Teache so ing and he pligh o u ban schools: A desc ip i e analysis. Educa ional e alua ion and policy analysis,24(1), 37–62. McFadden, D. (1974). The measu emen o u ban a el demand. Jou nal o public economics,3(4), 303–328. 15 Puga ch, T., & Sch oede , E. (2014). Incen i es o eache eloca ion: E idence om he gambian ha dship allowance. Economics o Educa ion Re iew,41(1), 120– 136. Ren, J., Xia, F., Chen, X., Liu, J., Hou, M., Shehzad, A., Sul ano a, N., & Kong, X. (2021). Ma ching algo i hms: Fundamen als, applica ions and challenges. IEEE T ansac ions on Eme ging Topics in Compu a ional In elligence,5(3), 332–350. Rosa, L. (2019). The o ganiza ion o educa ional ma ke s and effec s on indi iduals’ decisions: An empi ical analysis using b azilian educa ional policies.S an o d Uni e si y. 16 Figu e 1: Ma ke Conges ion - Numbe o applican s pe school 17 Table 1: Teache ans e s - F equency F eq. % o al % ask ans . % go ans . To al Numbe o Teache s 46755 - - - Teache s asking o ans e 8072 17.3% - - Teache s ans e ing 5804 12.4% 71.9% - T ans e s by p e e ence o de 1s 3057 - - 52.7% 2nd 810 - - 14% 3 h 382 - - 6.6% 4 h 230 - - 4% 5 h o mo e 1324 - - 22.8% No e: The able shows he o al numbe ( equency) o eache s egis e ed in he depa men o educa ion and he candida es o eache ans e s in he yea 2019. The sou ce o he da a is he adminis a i e eco ds om he depa men o educa ion in he ci y o Sao Paulo. 18 Table 2: Applican s - S a is ics (1) A e age numbe o schools lis ed 38.8 Median numbe o schools lis ed 7.0 % o eache s lis ing only 1 school 19.2 % o eache s lis ing only 2 schools 9.0 % o eache s lis ing only 3 schools 6.9 % o eache s lis ing only 4 schools 5.7 % o eache s lis ing 5 schools o mo e 59.1 No e: The able shows he s a is ics o applica ions. Fi s and second ows show he a e age and median numbe o schools lis ed by eache s in hei p e e ence lis . Thi d o se en h ows show he pe cen age o schools ha lis ed 1, 2, 3, 4, o mo e han 4 schools. The sou ce o he da a is he adminis a i e eco ds om he depa men o educa ion in he ci y o Sao Paulo. 19 Table 3: School cha ac e is ics by cu en and ans e ma ke Cu en T ans e Fi s Bes Assigned Tes sco e (z-sc ) -0.052 0.174 0.013 SES 5.05 5.133 5.082 Low SES (%) 23.626 20.364 22.241 Class size 28.73 28.821 28.804 No ice eache s (%) 0.246 0.165 0.209 School is in a ela 0.437 0.325 0.39 % o disabili y s uden s 0.025 0.026 0.025 Numbe class ooms 30.568 30.179 30.272 Re en ion a es 0.037 0.044 0.039 Dis home-school (km) 9.011 6.46 7.024 N 4807 4458 3481 No e: The able shows he school cha ac e is ic by school ype. Fi s column shows he cha - ac e is ics o he schools he eache we e cu en ly wo king. Second column shows he cha ac- e is ics lis ed as numbe one in he eache p e e ence o de . Thi d column shows he school cha ac e is ics o he school he eache s we e assigned. 20 Table 4: Rank-o de ed logi es ima es Teache p e e ences (1) (2) (3) (4) (5) (6) (7) Dis home-sch (log) 0.488∗∗∗ 0.489∗∗∗ 0.489∗∗∗ 0.490∗∗∗ 0.486∗∗∗ 0.488∗∗∗ 0.489∗∗∗ (0.019) (0.019) (0.019) (0.019) (0.019) (0.019) (0.019) Tes sco es (z-sco e) -0.095∗∗∗ -0.080∗∗∗ -0.072∗∗∗ -0.047∗∗ -0.056∗∗∗ -0.054∗∗∗ (0.013) (0.015) (0.015) (0.015) (0.015) (0.015) Pe c. low-ses 0.002 0.001 -0.000 0.001 0.001 (0.001) (0.001) (0.001) (0.001) (0.001) School in a ela 0.049∗∗ 0.047∗∗ 0.049∗∗ 0.047∗∗ (0.015) (0.015) (0.015) (0.015) Pe c. no ice eache s 0.725∗∗∗ 0.593∗∗∗ 0.589∗∗∗ (0.077) (0.082) (0.082) Numbe o class ooms -0.049∗∗∗ -0.050∗∗∗ (0.008) (0.008) Mone a y incen i e - Low 0.043 (0.027) Mone a y incen i e - High -0.016 (0.050) N40376 40376 40376 40376 40376 40376 40376 No e: The able shows es ima es om he ank-o de ed logi model. Nega i e signals indica e he school a ibu e is p e e ed by eache s ( eache s ank such schools fi s ), and posi i e signals indica e ha eache s dislike he a ibu e ( anking he schools lowe on hei p e e ence lis ). * p<0.05, ** p<0.01, *** p<0.001 21