89514                 Overview     Over  7  million  teachers  file  into  classrooms  across  Latin  America  and  the  Caribbean  each  day.  These   women  and  men  represent  4  percent  of  the  region’s  overall  labor  force  and  over  20  percent  of  its   technical  and  professional  workers.  Their  salaries  absorb  close  to  4  percent  of  the  region’s  GDP.  Their   working  conditions  vary  widely—from  mud-­‐walled,  one-­‐room  rural  schools  to  world-­‐class  facilities—but   Latin  America’s  teachers  share  one  important  commonality:  they  are  increasingly  recognized  as  the   critical  actors  in  the  region’s  efforts  to  improve  education  quality  and  results.   While  the  importance  of  good  teaching  may  be  intuitively  obvious,  only  over  the  past  decade  has   education  research  begun  to  quantify  the  high  economic  stakes  around  teacher  quality.  In  a  world   where  the  goals  of  national  education  systems  are  being  transformed,  from  a  focus  on  the  transmission   of  facts  and  memorization  to  a  focus  on  student  competencies—for  critical  thinking,  problem  solving   and  lifelong  learning—the  demands  on  teachers  are  more  complex  than  ever.  Governments  across  the   world  have  put  teacher  quality  and  teacher  performance  under  increasing  scrutiny.  The  Latin  America   and  the  Caribbean  (LAC)  region  is  no  exception  to  these  trends;  indeed,  in  some  key  areas  of  teacher   policy,  the  region  is  at  the  vanguard  of  global  reform  experience.     In  this  context,  this  study  aims  to:   • Benchmark  the  current  performance  of  LAC’s  teachers  and  identify  key  issues   • Share  emerging  evidence  on  important  reforms  of  teacher  policy  being  implemented  in  LAC   countries   • Analyze  the  political  “room  for  maneuver”  for  further  reform  in  LAC     We  focus  on  teachers  in  basic  education  (preschool,  primary,  and  secondary  education)  because  the   quantitative  and  qualitative  challenges  of  producing  effective  teachers  at  these  levels  differ  in  key  ways   from  university-­‐level  education,  which  has  been  addressed  in  other  recent  World  Bank  publications   (Rodriguez,  Dahlman,  and  Salmi  2008;  Salmi  2009).    We  also  focus  on  public  education  systems.     Notwithstanding  growing  basic  education  enrollments  in  private  schools  in  many  countries  in  the  region,   national  and  sub-­‐national  governments  deliver  the  bulk  of  basic  education  services  and  remain  the   guardians  of  education  quality  and  the  architects  of  education  policy.   Chapter  1  analyzes  global  and  regional  evidence  on  the  importance  of  education  results  for  economic   growth  and  competitiveness,  and  the  importance  of  teacher  quality  for  education  results.  It  profiles   LAC’s  teachers  and  how  their  characteristics  have  changed  in  recent  decades.  Chapter  2  provides  a  first-­‐ ever  insight  into  how  the  region’s  teachers  perform  inside  the  classroom,  drawing  on  new  research   conducted  for  this  report  in  over  15,000  classrooms  in  7  LAC  countries.     Chapters  3,  4,  and  5  focus  on  three  leading  areas  of  teacher  policy  reform  in  LAC  today:  chapter  3   analyzes  policies  to  recruit  better  teachers;  chapter  4  looks  at  programs  to  groom  teachers  and  improve   their  skills  once  they  are  in  service;  and  chapter  5  reviews  strategies  to  motivate  teachers  to  perform   their  best  throughout  their  career.     Chapter  6  analyzes  the  prominent  role  of  teachers’  unions  in  the  region  and  recent  country  experience   with  major  education  reforms.  This  chapter,  like  those  that  precede  it,  tries  to  distill  the  growing  body  of   evidence  from  within  and  outside  the  region  that  can  guide  the  design  of  effective  programs  and   sustainable  reforms.           Six  overarching  messages  emerge  from  this  study:     • The  low  average  quality  of  LAC’s  teachers  is  the  binding  constraint  on  the  region’s  education   progress,    and  consequently  on  the  contribution  of  national  education  spending  to  poverty   reduction  and  shared  prosperity   • Teacher  quality  in  the  region  is  compromised  by  weak  mastery  of  academic  content  as  well  as   ineffective  classroom  practice:  teachers  in  the  countries  studied  spend  65  percent  or  less  of  class   time  on  instruction  (compared  with  the  a  good  practice  benchmark  of  85  percent),  which  implies   the  loss  of  one  full  day  of  instruction  per  week;  they  make  limited  use  of  available  learning   materials,  especially  information  and  communications  technology  (ICT);  and  do  a  poor  job  of   keeping  students  engaged   • No  teaching  force  in  the  region  today  (except  possibly  Cuba’s)  can  be  considered  of  high  quality   against  global  comparators,  but  several  countries  have  made  progress  over  the  past  decade  in   raising  teacher  quality  and  student  learning  results,  most  notably  Chile   • There  are  three  fundamental  steps  to  a  high-­‐quality  teaching  force—recruiting,  grooming,  and   motivating  better  teachers—and  substantial  reform  experience  across  and  outside  of  LAC  in  all   three  areas  can  guide  the  design  of  better  policies   • Over  the  next  decade,  the  declining  size  of  the  school-­‐aged  population  in  about  half  of  the   countries  in  the  region,  notably  the  southern  cone,  could  make  it  substantially  easier  to  raise   teacher  quality;  in  the  other  half  of  the  region,  especially  Central  America,  the  need  for  more   teachers  will  complicate  the  challenge   • The  deepest  challenge  in  raising  teacher  quality  is  not  fiscal  or  technical,  but  political,  because   teachers’  unions  in  every  country  in  Latin  America  are  large  and  politically  active  stakeholders;   however,  a  growing  number  of  successful  reform  cases  is  yielding  lessons  that  can  aid  other   countries     Why  teachers  matter     LAC  education  performance  is  lagging.  Over  the  last  50  years,  Latin  American  and  Caribbean  countries   have  achieved  a  mass  expansion  of  education  coverage  that  took  a  century  or  more  to  accomplish  in   many  OECD  countries.  From  a  starting  point  of  less  than  10  percent  of  all  children  completing  secondary   school  in  1960,  today  most  LAC  countries  have  achieved  universal  primary  school  completion  and  high   rates  of  secondary  schooling.  Only  Guatemala  and  Haiti  stand  in  sharp  contrast  to  the  regional  progress.   While  the  average  LAC  worker’s  4  years  of  schooling    in  1960  was  little  more  than  half  the  level  of   workers  in  OECD  countries,  today  the  LAC  average  is  converging  on  the  OECD’s  average  of  12  years.   There  has  been  a  significant  and  rapid  accumulation  of  human  capital  in  Latin  America  over  the  past   half-­‐century.     But  recent  research  has  deepened  the  understanding  of  how  human  capital  contributes  to  economic   growth;  it  has  established  convincingly  that  what  counts  is  not  how  many  years  of  schooling  students   complete,  but  what  they  actually  learn.  It  may  seem  intuitively  obvious  that  a  year  of  schooling  in  Mali   will  not  equal  one  in  Singapore,  but  only  recently  have  researchers  been  able  to  quantify  this.  A  country   whose  average  performance  on  international  tests  is  one  standard  deviation  higher  than  another’s   (roughly  the  100-­‐point  difference  between  Mexico  and  Germany  on  the  2012  Program  for  International   Student  Assessment  [PISA]  exam)  will  enjoy  approximately  2-­‐percentage  point  higher  annual  long-­‐term   GDP  growth.  This  relationship  holds  across  countries  at  all  income  levels,  across  regions,  and  across   countries  within  regions  (Hanushek  and  Woessmann  2012,  figure  O.1).  Differences  in  countries’  average       level  of  cognitive  skills  are  consistently  and  fairly  strongly  correlated  with  long-­‐term  rates  of  economic   growth.  It  is  quality—in  terms  of  increased  student  learning—that  produces  the  economic  benefits  from   investing  in  education.  The  region’s  increasing  participation  in  international  and  regional  tests  provides   direct  evidence  of  how  well  its  students  are  learning,  and  four  important  conclusions  emerge.     Figure  O.1  Cognitive  skills  and  long-­‐term  economic  growth  across  regions       Source:  Hanushek  and  Woessmann  2012,  p.2.   Notes:  This  graph  compares  the  average  annual  rate  of  growth  (in  percent)  of  real  GDP  per  capita  in  1960–2000  (adjusted  for  the   initial  level  of  real  GDP  per  capita  in  1960)  with  average  scores  on  international  student  achievement  tests  over  this  period.     Region  codes:  Asia  (ASIA),  Commonwealth  OECD  members  (COMM),  Europe  (EURO),  Latin  America  (LATAM),  Middle  East  and   North  Africa  (MENA),  Sub-­‐  Saharan  Africa  (SSAFR).       First,  relative  to  its  level  of  economic  development,  the  LAC  underperforms  badly.  As  seen  in  figure  O.1,   LAC  countries’  average  learning  performance  on  all  international  tests  over  the  past  40  years  is  lower   than  that  of  every  other  region  except  Sub-­‐Saharan  Africa.  Of  the  65  countries  participating  in  the  2012   PISA  test,  all  eight  participating  LAC  countries  scored  below  the  average  for  their  level  of  per  capita   income  (figure  O.2).  The  nearly  100-­‐point  difference  between  the  Organisation  for  Economic  Co-­‐ operation  and  Development  (OECD)  average  math  score  (494)  and  the  average  for  participating  LAC   countries  (397)  represents  a  disparity  in  skills  equivalent  to  over  two  full  years  of  math  education.  The   gap  with  Shanghai,  whose  students  averaged  613,  is  more  than  five  years’  difference  in  math  skills.   Given  that  a  larger  share  of  all  15  year  olds  have  already  dropped  out  of  school  in  LAC  countries  than  in   the  OECD  or  East  Asia,  the  true  gap  in  skills  is  even  worse.  All  available  evidence  is  that  the  average   literacy  and  numeracy  skills  of  youths  in  LAC  badly  trail  those  of  other  middle-­‐income  countries.   Second,  the  range  in  performance  within  the  region  is  substantial.  Among  the  LAC  countries   participating  in  PISA,  the  gap  in  skills  between  the  top  performer  (Chile)  and  the  lowest  (Peru)  is  as  large   as  the  gap  between  Chile  and  Sweden  in  math  and  Chile  and  the  United  States  in  reading.  Regional  tests   show  that  LAC  countries  that  do  not  participate  in  PISA  are  even  further  behind:  countries  such  as   Honduras,  Venezuela,  and  Bolivia  are  very  far  off  track  in  terms  of  the  amount  of  globally  relevant   learning  a  year  of  schooling  produces.       Figure  O.2  PISA  reading  scores  and  income  per  capita  for  LAC  countries,  2012   550 500 Costa RicaChile 450 Mexico Brazil Uruguay 400 Colombia Argentina Peru 350 0 10 20 30 40 50 60 70 GDP per capita, US$ thousands   Source:  OECD/PISA  2013  and  World  Bank.   Note:  GDP  per  capita  is  adjusted  for  purchasing  power  parity  (PPP),  in  2005  constant  prices.           The  third  broad  conclusion  is  encouraging:  some  LAC  countries  are  making  sustained  progress  in  closing   the  gap  with  the  OECD.  Between  2000  and  2012,  Chile,  Brazil,  and  Peru  registered  some  of  the  biggest   improvements  in  the  entire  PISA  sample:  more  than  twice  the  annual  rate  of  improvement  seen  in  the   United  States  or  Korea  (Hanushek,  Peterson,  and  Woessmann  2012).1  While  the  LAC  region  as  a  whole   faces  a  large  challenge,  there  is  also  the  encouraging  prospect  of  relevant  lessons  from  within  the   region.     Figure  O.3  Comparative  PISA  math  improvement  for  LAC  countries,  2000–12   500   Brazil   Argentina   470   Chile   440   Colombia   410   Peru   380   Uruguay   350   Mexico   320   Costa  Rica   290   OECD  Average   2000   2003   2006   2009   2012     Source:  OECD  PISA  Secretariat  2013.   But  the  final  conclusion,  reinforced  by  the  2012  PISA  results,  is  that  there  is  no  room  for  complacency.   LAC’s  top  improvers  as  well  as  other  countries  in  the  region  all  made  far  less  progress  than  in  prior   rounds  of  PISA.  This  suggests  that  while  other  countries  in  the  region  may  have  something  to  learn  from       the  policies  adopted  in  Chile,  Brazil,  and  Peru  over  the  past  decade,  those  countries  also  need  to  do   more.     Teacher  quality  drives  learning.  If  the  economic  benefits  from  education  investments  hinge  on  their   effectiveness  in  producing  student  learning,  the  critical  question  becomes,  what  drives  learning?”   Students’  family  background  (parent  education,  socioeconomic  status,  and  conditions  at  home  such  as   access  to  books)  remains  the  largest  overall  predictor  of  learning  outcomes.  A  growing  body  of  research   confirms  the  importance  of  policies  to  protect  children’s  nutrition,  health,  cognitive,  and  socioemotional   development  in  the  earliest  years  of  life.  But  research  over  the  past  decade  has  also  built  new  evidence   that  once  children  get  to  school,  no  single  factor  is  as  critical  as  the  quality  of  teachers.     Increasingly  abundant  student  test  data,  especially  in  the  United  States,  allow  researchers  to  measure   the  “value  added”  of  individual  teachers  over  the  course  of  a  single  school  year,  and  has  generated  eye-­‐ opening  evidence  of  widely  varying  teacher  effectiveness,  even  within  the  same  school  and  same  grade.   Students  with  a  weak  teacher  may  master  50  percent  or  less  of  the  curriculum  for  that  grade;  students   with  a  good  teacher  get  an  average  gain  of  one  year;  and  students  with  great  teachers  advance  1.5   grade  levels  or  more  (Hanushek  and  Rivkin  2010;  Rockoff  2004).  The  most  recent  research  shows  that   exposure  to  even  a  single  highly  effective  teacher  raises  a  student’s  college  participation  rates  and   subsequent  income  (Chetty  et  al.  2014).  A  series  of  great  or  bad  teachers  over  several  years  compounds   these  effects  and  can  lead  to  unbridgeable  gaps  in  student  learning  levels.  No  other  attribute  of  schools   comes  close  to  this  impact  on  student  achievement.     This  evidence  has  intensified  the  focus  of  policymakers  and  researchers  on  how  to  identify  effective   teachers.  No  one  disputes  that  a  comprehensive  definition  of  teacher  quality  must  encompass  many   different  teacher  characteristics  and  performance  dimensions.  But  teachers’  ability  to  ensure  that  their   students  learn  is  the  sine  qua  non  for  students  and  nations  to  reap  education’s  economic  and  social   benefits.   Who  are  LAC’s  teachers?   What  do  we  know  about  the  characteristics  and  performance  of  teachers  in  Latin  America?  Available   data  paint  a  distressing  picture.   Mostly  female,  with  relatively  low  socioeconomic  status.  About  75  percent  of  Latin  America’s  teachers   are  female,  but  this  ranges  from  a  low  62  percent  in  Mexico  to  82  percent  in  Uruguay,  Brazil,  and  Chile.   Teachers  are  also  poorer  than  the  overall  pool  of  university  students.  University  entrance  data  show  that   students  majoring  in  education  are  of  lower  socioeconomic  status  and  are  more  likely  to  be  first-­‐ generation  university  students  than  entrants  in  other  fields;  the  data  point  to  a  pool  of  students  whose   lives  may  have  afforded  them  limited  experience  with  other  professions  and,  consequently,  more   limited  academic  aspirations.  The  teaching  force  in  most  of  Latin  America  is  also  aging.  In  Peru,  Panama,   and  Uruguay,  the  average  teacher  is  more  than  40  years  old;  the  youngest  corps  in  the  region,  in   Honduras  and  Nicaragua,  average  35  years  of  age.   High  levels  of  formal  education,  but  weak  cognitive  skills.  Teachers’  formal  education  has  continued  to   rise  across  the  LAC  region.  In  1995,  only  19  percent  of  Brazilian  primary  teachers  had  university  degrees;   in  2010,  62  percent  did.  In  all  ten  LAC  countries  for  which  comparable  household  survey  data  are   available,  the  formal  educational  level  of  teachers  today  is  higher  than  for  all  other  professional  and   technical  workers  and  considerably  higher  than  for  office  workers.   The  increase  in  formal  education,  however,  is  undercut  by  evidence  that  the  individuals  entering   teaching  in  Latin  America  are  academically  weaker  than  the  overall  pool  of  higher  education  students.   Fifteen-­‐year-­‐old  students  who  identify  themselves  as  interested  in  a  teaching  career  have  much  lower       PISA  math  scores  than  students  interested  in  engineering  in  every  country  in  the  region,  and  score   below  the  national  average  in  every  country  except  Uruguay  (figure  O.4).     Data  from  university  entrance  exams  paint  a  similar  picture.  Students  applying  to  teacher  education   programs  average  505  on  the  Chilean  university  entrance  exam  (PSU);  the  average  for  law  is  660;   engineering,  700;  and  medicine,  745.  At  the  University  of  Sao  Paulo,  students  applying  for  law  and   engineering  programs  score  36  percent  higher  than  teacher  education  applicants,  and  medical  school   applicants  score  50  percent  higher.     Figure  O.4  Comparative  PISA  math  performance  of  prospective  teachers  and  prospective  engineers   580   530   480   430   380   330   280   bra*   chi   ury   isr   ita   grc   prt   hrv   usa   esp   nor   lux   gbr   swe   aut   est   aus   che   nzl   kor   lie   fra   mex   arg   deu   ind   cze   jpn   tha   isl   hkg   irl   pol   bel   rou   svn   can   kgz   twn   hun   svk   dnk   National  Average  Score   Average  Score  Prospective  Engineers   Average  Score  Prospective  Teachers   National  Average  Score  LAC  Countries     Source:  OECD,  PISA,  2000–06.  Data  are  from  PISA  2006,  except  for  Brazil  (from  PISA  2000).   There  are  few  direct  studies  of  how  much  Latin  American  teachers  know  about  the  subjects  they  teach,   but  those  available  show  a  disturbing  disconnect  between  teachers’  formal  credentials  and  their   cognitive  skills.  Fully  84  percent  of  sixth-­‐grade  teachers  in  Peru  scored  below  level  2  on  a  2006  test   where  level  3  meant  mastery  of  sixth-­‐grade  math  skills.  On  tests  of  teacher  content  mastery  in   Colombia,  Ecuador,  and  Chile,  fewer  than  3  percent  of  teachers  have  scored  in  the  range  considered   excellent.     On  the  one  international  study  that  directly  compares  teachers’  mastery  of  math,  Chile  was  the  sole  LAC   country  to  participate.  The  study  tested  the  math  skills  of  teacher  education  students.  Chile’s  future   secondary  school  math  teachers  scored  the  lowest  of  the  participating  countries,  and  its  future  primary   school  teachers  were  second  lowest  (figure  O.5).  Most  of  the  countries  in  the  study  were  high-­‐income,   high-­‐achieving  countries.  Yet  Chile’s  future  secondary  school  teachers  had  weaker  math  skills  than   teachers  from  Botswana  and  the  Philippines.  Among  future  primary  school  teachers,  only  those  from   Georgia  performed  worse.  Given  that  Chile  is  the  LAC  region’s  highest  performer  on  international  tests,   these  data  point  to  deep  issues  for  other  countries  in  the  region  in  raising  the  expertise  of  their   teachers.             Figure  O.5  Comparative  math  content  knowledge  of  future  math  teachers,  2008   a.  Primary  school  teachers   b.  Secondary  school  teachers   Taipei,  Taiwan,  China   623   Singapore   544   Singapore   586   Switzerland   531   Switzerland   548   Russian  Federalon   536   Poland   529   United  States   518   Germany   483   Norway   509   United  States   468   Germany   501   Norway   461   Spain   481   Poland   456   Philippines   442   Botswana   441   Botswana   436   Philippines   440   Chile   354   Chile   413   0   200   400   600   Georgia   345     0   200   400   600   800     Source:  TEDS-­‐M  2008.       Relatively  low  salaries.  What  explains  relatively  weak  students  pursuing  high  levels  of  formal  education   to  become  teachers?  What  are  the  incentives  to  enter  teaching  today  in  LAC?  Figure  O.6  compares   salaries  for  teachers  to  those  of  all  other  professional  workers  in  all  ten  of  the  countries  for  which   comparable  household  survey  data  are  available  controlling  for  age,  gender,  urban/rural  location,  and   employment  experience.  On  a  monthly  basis,  teachers’  salaries  in  2010  were  between  10  and  50   percent  lower  than  salaries  for  other  “equivalent”  professional  workers,  and  have  been  throughout  the   2000s.     Teachers,  however,  work  significantly  fewer  hours,  reporting  30  to  40  hours  per  week  on  average,   compared  with  40  to  50  hours  per  week  for  other  professional,  technical,  and  office  workers.  Adjusted   for  working  hours,  teachers’  relative  position  is  different.  In  three  countries  (Mexico,  Honduras,  and  El   Salvador)  teachers  earn  20  to  40  percent  more  than  comparable  professional  and  technical  workers;  in   three  others  they  are  on  par  (Costa  Rica,  Uruguay,  and  Chile);  and  in  four  countries  they  earn  10  to  30   percent  less  (Peru,  Panama,  Brazil,  and  Nicaragua).                     Figure  O.6  Average  salaries  for  teachers  and  other  professional  workers,  2000  and  2010   a. Monthly  salaries   b. Salaries  adjusted  for  working  hours   percent   percent   35%   15%   5%   25%   -­‐5%   15%   -­‐15%   5%   -­‐25%   -­‐35%   -­‐5%   -­‐45%   -­‐15%   -­‐55%   -­‐25%   -­‐65%   -­‐35%   Late  2000s   Early  2000s   Late  2000s   Early  2000s       Source:  World  Bank  analysis  of  household  survey  and  labor  market  data  for  10  LAC  countries.     Note:  All  values  control  for  age,  education,  gender,  and  urban/rural  location  and  are  adjusted  for  inflation.  Teacher  income  is   income  from  teaching  only,  but  can  reflect  multiple  teaching  jobs.   Flat  salary  trajectory.  Underlying  these  differences  in  average  salaries,  however,  is  a  much  flatter   lifetime  career  trajectory  for  teachers  than  for  other  professional,  technical,  and  office  workers.   Teachers’  entering  salaries  in  LAC  are  on  par  with  other  professional  and  technical  workers  in  many   countries  but  diverge  significantly  thereafter.  Teachers’  salaries  rise  very  slowly,  while  other  workers   reap  salary  gains  as  their  experience  increases.  There  is  also  little  wage  differentiation  in  education   compared  with  other  sectors:  irrespective  of  individual  skills,  talent,  and  experience,  landing  a  job  in   teaching  guarantees  a  salary  within  a  relatively  narrow  band,  with  little  risk  of  a  very  low  or  declining   wage  but  little  chance  of  a  high  one.     Research  by  Hernani-­‐Limarino  (2005)  suggests  that  in  Latin  America  and  the  Caribbean  individuals  who   tend  to  be  less  productive  earn  relatively  more  as  teachers  and  those  with  attributes  that  make  them   highly  productive  tend  to  earn  relatively  less.  This  echoes  research  by  Hoxby  and  Leigh  (2004)  for  the   United  States  which  finds  that  talented  women  have  been  driven  out  of  teaching  over  the  past  several   decades  by  the  “push”  of  a  highly  compressed  wage  scale,  which  is  unattractive  to  more  ambitious   individuals.  It  is  not  only  in  the  United  States  and  Latin  America  that  the  average  cognitive  ability  of   university  students  getting  teaching  degrees  has  declined  and  higher-­‐ability  individuals  have  left  the   profession  at  a  higher  rate.  Researchers  in  Sweden  have  documented  similar  trends  (Corcoran  et  al.   2004;  Eide  et  al.  2004;  Hoxby  2004;  Fredriksson  and  Ockert  2007).     Job  stability.  As  Mizala  and  Nopo  (2011)  have  observed,  many  nonpecuniary  or  “intrinsic”  attributes  of   the  teaching  profession  compensate  for  its  relatively  weak  salary  incentives  in  the  eyes  of  prospective   teachers.  These  include  the  mission  of  helping  children  and  the  satisfactions  of  professional  mastery  and   collegial  interaction.  Teaching  also  offers  long  vacations,  relatively  generous  health  and  pension   benefits,  and  a  “family-­‐friendly”  short  official  working  day.  Perhaps  the  most    powerful  attraction  is  high   job  security.  Labor  market  data  show  that  for  women  in  particular,  teaching  offers  stable  employment;   women  who  have  graduated  from  teacher  education  over  the  past  40  years  are  significantly  more  likely   to  have  been  employed  and  stayed  employed  than  women  with  other  degrees.         Excess  supply.  Tertiary-­‐level  teacher  education  programs  have  proliferated  in  LAC  over  the  past  15   years.  The  costs  of  establishing  such  programs  are  low,  and  have  attracted  a  large  number  of  private   providers  into  the  field.  From  the  demand  side,  the  low  or  nonexistent  academic  standards  for  entry  to   these  programs  make  them  attractive  to  a  rapidly  expanding  pool  of  secondary  school  graduates.   Virtually  all  countries  in  the  region  report  difficulty  finding  sufficient  teachers  for  specialty  subjects  such   as  secondary  school  math  and  science,  or  for  bilingual  schools  in  rural  areas.  But  the  broader  picture   across  the  region  today  is  substantial  excess  production  of  teacher  graduates  of  generally  low  academic   quality.  Recent  data  for  Peru,  Chile,  Costa  Rica,  Panama,  and  Uruguay  suggest  that  40  to  50  percent  of   graduates  from  teacher  training  schools  will  not  find  work  as  teachers.     Available  evidence  suggests  that  Latin  America  is  not  attracting  the  high-­‐caliber  individuals  it  needs  to   build  world-­‐class  education  systems.  Virtually  all  countries  in  the  region  appear  trapped  in  a  low-­‐level   equilibrium  of  low  standards  for  entry  into  teaching,  low-­‐quality  candidates,  relatively  low  and   undifferentiated  salaries,  low  professionalism  in  the  classroom,  and  poor  education  results.  Moving  to  a   new  equilibrium  will  be  difficult.  No  Latin  American  school  system  today,  except  possibly  Cuba’s,  is  very   close  to  high  standards,  high  academic  talent,  high  or  at  least  adequate  compensation,  and  high   professional  autonomy  that  characterize  the  world’s  most  effective  education  systems  (such  as  those   found  in  Finland;  Singapore;  Shanghai,  China;  Korea;  Switzerland;  the  Netherlands;  and  Canada.     LAC’s  teachers  inside  the  classroom     The  magic  of  education—the  transformation  of  schooling  inputs  into  learning  outcomes—happens  in   the  classroom.  Every  element  of  an  education  system’s  expenditure,  from  curriculum  design  through   school  construction,  book  procurement,  and  teacher  salaries,  comes  together  at  the  moment  when  a   teacher  interacts  with  students  in  the  classroom.  How  intensively  this  instructional  time  is  used  is  a  core   determinant  of  the  productivity  of  education  spending.     Research  conducted  for  this  study  provides  a  first-­‐ever  look  inside  LAC’s  classrooms  to  examine  how   teachers  use  class  time  and  other  available  resources  to  support  their  students’  learning.  Over  15,000   classrooms  in  more  than  2,700  schools  in  seven  different  countries  were  observed  between  2009  and   2012:  the  largest-­‐scale  international  study  of  this  kind  ever  mounted.  Through  unannounced  visits  to   national  (or  state-­‐level)  representative  samples  of  schools,  trained  observers  used  a  standardized   research  protocol  called  the  “Stallings  classroom  snapshot”  to  generate  internationally  comparable  data   on  four  variables:     • Teachers’  use  of  instructional  time   • Teachers’  use  of  materials,  including  computers  and  other  ICT   • Teachers’  core  pedagogical  practices   • Teachers’  ability  to  keep  students  engaged   The  method  was  originally  developed  in  the  United  States,  so  LAC  results  may  be  benchmarked  against   data  from  US  school  systems  collected  over  several  decades  by  researchers  Stallings  and  Knight  (2003).   Evidence  from  the  observations  in  LAC  supports  five  main  conclusions:     Low  use  of  instructional  time  contributes  to  low  student  learning  in  LAC.  No  school  system  in  LAC   studied,  either  at  the  national  or  state  level,  comes  close  to  the  Stallings  good  practice  benchmark  of  85   percent  of  total  class  time  used  for  instruction.  The  highest  averages  recorded—65  percent  for  the   national  sample  in  Colombia,  and  64  percent  for  Brazil  and  Honduras—are  a  full  20  percentage  points   below  what  Stallings’  research  suggested  a  well-­‐run  classroom  achieves  (figure  O.7).  Since  Stallings   measures  are  statistically  representative  of  the  functioning  of  the  school  system  as  a  whole,  this  implies       that  20  percent  of  potential  instructional  time  is  being  lost  across  Latin  America  compared  with  the  good   practice  goal.  This  is  the  equivalent  of  one  less  day  of  instruction  per  week.     Figure  O.7  Average  time  on  instruction  in  LAC  countries     100%   85%   Percent  of  Total  Class  Time   80%   64%   64%   65%   61%   62%   60%   52%   39%   40%   28%   25%   24%   27%   25%   14%   12%   15%   20%   9%   11%   10%   9%   0%   0%   Mexico  D.F.   Jamaica   Peru   Honduras   Brazil   Colombia   Stallings   Good   Practice   Indicator   Academic  Activities   Classroom  Management   Teacher  Off-­‐Task     Note:  Results  for  the  Dominican  Republic  are  not  included  because  the  sample  was  a  pilot.     Source:  World  Bank  classroom  observation  database   Most  of  the  time  lost  to  instruction  is  used  on  classroom  management  activities,  such  as  taking   attendance,  cleaning  the  blackboard,  grading  homework,  or  distributing  papers,  which  absorb  between   24  percent  and  39  percent  of  total  class  time:  well  above  the  15  percent  benchmark.  Teacher  training   programs  in  many  OECD  countries  impart  techniques  for  managing  classroom  transitions  and   administrative  processes  as  efficiently  as  possible,  with  the  mantra  that  “instructional  time  is  a  school’s   most  expensive  resource.”  Classroom  teachers  in  Latin  America  appear  to  operate  with  little  of  this   pressure.     Stallings  benchmarks  also  assume  that  teachers  spend  the  entire  class  session  either  teaching  or   managing  the  classroom,  but  in  every  LAC  country  studied  teachers  spend  at  least  9  percent  of  time   engaged  in  neither  of  these,  which  is  considered  teacher  time  completely  “off-­‐task”  (figure  O.8).  The   highest  shares  are  13  percent  in  Peru,  12  percent  in  Honduras,  and  11  percent  in  Jamaica.  In  some   systems,  teachers  are  physically  absent  from  the  room  as  much  as  6  to  11  percent  of  total  class  time.  In   others,  as  much  as  6  to  8  percent  of  the  time  teachers  are  engaged  in  social  interaction  with  someone  at   the  classroom  door  or  simply  not  interacting  with  the  class.  Ten  percent  of  total  instructional  time  off-­‐ task  equals  20  lost  days  in  a  200-­‐day  school  year.  In  these  countries,  more  than  half  of  the  lost  days  of   instruction  are  because  teachers  are  physically  absent  from  the  classroom,  arriving  late  to  class,  leaving   early,  or  conducting  other  school  business  during  class  time.                   Figure  O.8  Breakdown  of  teacher  time  off-­‐task,  by  country   14%   Percent  of  total  class  time   12%   10%   3.8%   8%   4.1%   5.0%   6.0%   11.1%   6%   5.8%   4%   7.5%   4.0%   3.0%   2.4%   4.0%   2%   1.4%   1.3%   2.0%   2.5%   0%   0.7%   0.6%   Mexico  D.F.   Colombia   Brazil   Jamaica   Honduras   Peru   Teacher  Social  Interaction  with  Students   Teacher  Social  Interaction  or  Uninvolved     Teacher  out  of  the  Room     Source:  World  Bank  classroom  observation  database.   In  Brazil,  Honduras,  Mexico’s  D.F.,  and  Colombia,  student  test  data  permit  correlation  of  teachers’  use  of   time  with  learning  results  at  the  school  level.  Table  O.1  shows  a  characteristic  result:  a  very  different   pattern  of  time  use  between  the  highest-­‐  and  lowest-­‐performing  schools  in  Rio  de  Janeiro  (on  a   combined  index  of  student  test  scores  and  student  pass  rates  called  the  Index  of  Basic  Education   Development  [IDEB]).  Top  schools  averaged  70  percent  of  class  time  on  instruction  and  27  percent  of   time  on  classroom  management.  Teachers  were  off-­‐task  only  3  percent  of  the  time  and  were  never   absent  from  the  classroom.  In  the  lowest-­‐performing  schools,  only  54  percent  of  time  was  spent  on   instruction,  39  percent  was  absorbed  in  classroom  management,  and  teachers  were  off-­‐task  7  percent   and  physically  absent  3  percent  of  the  time.  These  data  mean  that  students  in  Rio’s  high-­‐performing   schools  receive  an  average  of  32  more  days  of  instruction  over  the  200-­‐day  school  year  than  their   counterparts  in  low-­‐performing  schools.  Observation  data  cannot  establish  causality,  but  gaps  of  this   magnitude  in  opportunities  to  learn  could  clearly  contribute  to  gaps  in  test  scores  and  pass  rates.     Table  O.1  Use  of  class  time  in  Rio  de  Janeiro  schools,  2010   percent  of  total  class  time   Use  of  class  time     Teacher  out  of   Fifth  grade   Time  on   Classroom   Teacher  off-­‐task   classroom  (within   instruction   organization   off-­‐task)   Rio  municipality   58%   37%   6%   1%   Top  10%  of  schools  on   70%   27%   3%   0%   IDEB   Bottom  10%  of  schools   54%   39%   7%   3%   on  IDEB     Difference   0.16   -­‐0.13   -­‐0.03   -­‐0.03       [0,09]*   [0,09]*   [0,02]   [0,01]**   Note:  Robust  standard  errors  in  parentheses:  *statistically  significant  at  10  percent  level,  **statistically  significant  at   5  percent  level.  .   Source:    World  Bank  classroom  observation  database.         Schools  at  the  top  and  bottom  ends  of  the  performance  distribution  in  terms  of  student  learning  almost   always  show  large,  statistically  significant  differences  in  instructional  time.  In  Honduras,  the  top  10   percent  of  schools  on  the  national  assessment  averaged  68  percent  of  time  on  instruction,  while  the   schools  in  the  bottom  10  percent  averaged  46  percent.  In  Mexico’s  Federal  District,  the  10  percent  of   schools  with  the  highest  test  scores  averaged  62  percent  of  time  on  instruction  compared  with  51   percent  of  time  on  instruction  in  the  10  percent  of  schools  with  the  lowest  scores.  Across  the  full   distribution,  positive  correlations  hold  in  all  four  countries,  across  all  tested  grades  and  subjects,  with   relatively  few  exceptions:  students  have  higher  learning  outcomes  in  schools  where  teachers  devote   more  time  to  academic  activities.     When  test  scores  are  correlated  with  teachers’  ability  to  keep  students  engaged  in  what  they  are   teaching,  the  results  can  be  even  stronger.  Teachers  who  successfully  involve  the  entire  class  in  the  task   at  hand  have  better  control  of  the  class,  fewer  problems  with  discipline,  more  time  to  impact  student   learning,  and  afford  a  larger  share  of  their  students  the  opportunity  to  learn  (figure  O.9).   Figure  O.9  Time  on  instruction  with  high  student  engagement  and  student  learning  outcomes   a.  Colombia  fifth-­‐grade  math   b.  Honduras  third-­‐grade  language       Source:  World  Bank  classroom  observation  database.     Strongest  of  all  was  the  negative  correlation  between  teachers’  time  off-­‐task  and  student  achievement.   As  discussed  earlier,  approximately  10  percent  of  the  time  LAC  teachers  are  neither  teaching  nor   managing  the  classroom.  This  pattern  has  clear  consequences  for  student  learning  (figure  O.10).   Although  broad  patterns  are  apparent,  the  correlations  are  statistically  weaker  than  might  be  expected.   A  major  explanation  is  that  student  test  score  data  are  analyzed  at  the  school  level  and  represent  the   average  score  for  several  different  classrooms  of  students  (for  example,  three  or  four  different  fourth-­‐ grade  math  classes)  in  a  given  school,  while  the  observation  data  are  for  only  one  of  those  classrooms.   Given  large  variation  in  classroom  dynamics  from  one  teacher  to  another,  even  in  the  same  school—as   discussed  later  in  this  section—comparing  school-­‐level  average  learning  outcomes  with  the  dynamics  of   a  single  classroom  injected  a  large  degree  of  random  variation  into  the  correlations.                 Figure  O.10  Teacher  time  off-­‐task  and  student  learning  in  LAC  countries   a.  Colombia  fifth-­‐grade  Spanish  (SABER)   b.  Honduras  sixth-­‐grade  language       c.  Mexico  ninth-­‐grade  math  (ENLACE)   d.  Rio  fifth-­‐grade  math  (PROVARIO)     Time Off Task and PROVARIO 5 Math Score 2 1 PROVARIO 5 Math -Std 0 -1 -2 0 .1 .2 .3 .4     Time Off Task   Teachers  rely  heavily  on  the  blackboard  and  make  little  use  of  ICT.  A  second  finding  of  the  research  is   that  many  learning  materials  available  in  LAC  classrooms  are  not  used  intensively  by  teachers.   Descriptive  data  collected  by  the  observers  shows  that  most  schools  in  the  region  today  offer  students  a   reasonably  enriched  learning  environment.  Students  are  widely  equipped  with  workbooks,  writing   materials,  and  textbooks.  A  fast-­‐growing  share  of  schools  has  visible  ICT  in  the  classroom:  from   televisions  to  digital  whiteboards,  LCD  projectors,  and  laptops.  In  this  sample,  both  Peru  and  Honduras   had  introduced  one  laptop  per  child  (OLPC).     But  teacher  practice  continues  to  rely  heavily  on  a  single,  very  traditional,  learning  aid:  the  blackboard.   About  one-­‐third  of  all  time  spent  on  teaching  activities,  teachers  use  the  blackboard  and  nothing  else.   Between  14  and  24  percent  of  the  time,  teachers  use  no  learning  materials.  Teachers  use  available   classroom-­‐level  ICT  only  2  percent  of  the  time.  In  Honduras  and  Peru—the  countries  with  the  largest   investments  in  one-­‐to-­‐one  computing  in  this  sample—the  share  of  total  class  time  spent  using  these   materials  was  in  fact  the  lowest,  1  percent  in  Peru  (in  2011)  and  less  than  1  percent  in  Honduras  (in   2011).     Students  are  unengaged.  LAC’s  teachers  have  great  difficulty  keeping  their  students  engaged  in   learning.  In  no  system  studied  do  teachers  on  average  keep  the  entire  class  engaged  in  learning  more   than  25  percent  of  class  time  (figure  O.11).  More  than  half  of  all  class  time,  in  all  countries,  up  to  five   students  are  tuned  out.  Between  one-­‐fifth  and  one-­‐quarter  of  total  class  time,  in  all  countries,  a  large   group  of  students  (six  or  more)  is  visibly  not  involved  in  the  activity  the  teacher  is  leading.  With  an   average  class  size  of  25  across  our  sample,  six  students  represent  a  significant  share,  and  can  disrupt  the       work  of  other  students.  Observers  in  every  country  saw  classrooms  that  were  badly  out  of  control,  even   with  the  teacher  present  and  aware  of  being  observed.       Figure  O.11  Teacher  time  on  instruction  with  the  entire  class  engaged   70%   64%   64%   65%   61%   62%   Percent  of  Total  Class  TIme   60%   52%   50%   40%   30%   25%   23%   21%   21%   19%   19%   20%   10%   0%   Mexico  D.F.   Jamaica   Peru   Honduras   Brazil   Colombia   Time  on  Instruclon   Time  on  Instruclon  with  Enlre  Class  Engaged     Source:    World  Bank  classroom  observation  database.     Given  high  repetition  rates,  LAC’s  teachers  often  face  classes  that  span  different  ages  and  learning   levels.  Well-­‐trained  teachers,  however,  learn  to  handle  such  classes  and  with  well-­‐designed  lesson  plans   keep  all  students  engaged  much  more  than  20  percent  of  the  time.  One  of  the  clearest  findings  of  this   research  is  that  poor  student  learning  results  can  be  directly  linked  to  the  failure  of  teachers  to  keep   students  engaged  in  learning.  Important  challenges  for  both  pre-­‐service  and  in-­‐service  teacher  training   programs  in  LAC  are  to  ensure  that  teachers  recognize  the  importance  of  drawing  all  students  into  the   learning  process,  are  equipped  with  a  range  of  teaching  strategies  to  achieve  this,  and  arrive  at  school   each  day  prepared  to  use  these  strategies,  and  every  minute  of  class  time,  effectively.       Average  classroom  practice  varies  tremendously  across  schools.  A  fourth  finding  is  the  wide  variation   in  average  classroom  practice  across  schools.  In  every  system,  there  are  many  schools  where  the   average  share  of  class  time  used  for  instruction  exceeds  the  Stallings  85  percent  benchmark  and  others   where  instructional  time  is  disastrously  low:  below  20  percent  of  total  class  time.  Imagine  attending  a   school  where  four  days  per  week  there  is  no  instruction  (figure  O.12).   In  some  countries,  there  were  distinct  differences  across  regions;  in  Honduras,  schools  in  the  province  of   Colon  averaged  33  percent  of  time  on  instruction  while  those  in  Copan  averaged  83  percent.  In   Pernambuco,  Brazil,  schools  observed  in  late  2009  that  went  on  to  achieve  their  performance  targets  for   the  year  and  earn  a  school-­‐level  bonus  averaged  63  percent  of  time  on  instruction  compared  to  54   percent  in  schools  that  failed  to  gain  the  bonus.               Figure  O.12  Distribution  of  schools  by  average  time  spent  on  instruction   a. Colombia  across  school  variation   b.  Peru  across  school  variation   Colombia Across School Variation Peru AcrossSchool Variation 4 2 3 1.5 Density Density 2 1 .5 1 0 0 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .2 .4 .6 .8 1 Time on Instruction Time on Instruction   kernel = epanechnikov, bandwidth = 0.0503   Honduras Across School Variation Jamaica Across School Variation 4 4 3 3 Density Density 2 2 1 1 0 0 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 Time on Instruction Time on Instruction     Mexico Across School Variation Brazil Across School Variation 4 4 3 3 Density Density 2 2 1 1 0 0 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 Time on Instruction Time on Instruction     Source:  World  Bank  classroom  observation  database     What  is  evident  from  this  dispersion  is  that  school  systems  are  not  focused  on  the  issue  of  instructional   time.  Within  a  given  national  or  subnational  education  system,  schools  operate  in  the  same  institutional   and  policy  environment.  They  share  the  same  policies  for  the  selection  of  school  directors,  curriculum,   teacher  standards,  preparation,  and  student  assignment  rules.  Yet  this  research  shows  that  these   policies  are  playing  out  at  the  school  level  in  widely  different  ways.  And  these  difference  have  important       consequences  for  students.  In  Honduran  schools,  64  percent  of  time  is  spent  on  instruction,  on  average,   across  the  system.  But  the  top  20  percent  of  schools  (in  terms  of  instructional  time  use)  average  85   percent:  they  achieve  the  Stallings  benchmark.  Schools  in  the  bottom  quintile  average  37  percent.  This   48  percentage-­‐point  difference  means  that  students  in  top  quintile  schools  receive  an  average  of  96   more  days  of  instruction  than  students  in  bottom  schools  each  year.  The  consequences  for  students   spending  several  years  in  a  school  that  consistently  delivers  much  less  instruction  are  cumulative  and   highly  unfair.     Average  classroom  practice  varies  tremendously  within  schools.  The  fifth  and  most  surprising  finding  is   the  degree  of  variation  in  teacher  practice  inside  schools.  The  difference  in  instructional  time  use   between  the  best  (highest  time  on  teaching)  and  worst  teacher  in  a  single  school  is  typically  about  two-­‐ thirds  as  large  as  the  variation  across  the  whole  school  system,  a  truly  staggering  range.     Figure  O.13  draws  out  these  patterns.  In  Colombia,  for  example,  the  average  time  on  instruction  across   the  national  sample  is  65  percent.  Schools  in  the  top  quintile  of  the  performance  distribution  (of  time   use)  average  82  percent,  and  schools  in  the  bottom  quintile  average  49  percent:  a  33  percentage-­‐point   difference.  But  an  even  greater  range  exists  between  the  best-­‐  and  worst-­‐performing  teachers  inside   these  schools.  For  a  school  in  the  lowest  quintile,  the  best  teacher  will  typically  spend  78  percent  of  class   time  on  instruction—not  far  from  the  Stallings  benchmark—while  the  worst  teacher  in  that  school  will   use  only  18  percent  of  time  for  instruction.  This  60  percentage-­‐point  average  difference  in  time  spent   teaching  within  these  schools  is  larger  than  the  range  in  the  average  across  quintiles  of  schools.     Figure  O.13  Colombia  and  Honduras  variation  in  instructional  time  within  schools,  2011   Percent  of  class  time  spent  on  instruction   a.  Colombia   b.  Honduras   100%   100%   80%   80%   60%   60%   40%   40%   20%   20%   0%   Quintile   Quintile   Quintile   Quintile   Quintile   0%   1   2   3   4   5   Quintile   Quintile   Quintile   Quintile   Quintile   1   2   3   4   5   Lowest  Classroom  on  Task   Lowest  Classroom  on  Task   Average  Time  on  Task  in  School   Average  time  on  Task  in  School   Highest  on  Task  Classroom   Highest  on  Task  Classroom       Source:  World  Bank  classroom  observation  database.       Note:  Quintiles  refer  to  the  distribution  of  schools  ranked  from  highest  to  lowest  average  time  on  instruction.       In  the  20  percent  of  schools  with  the  highest  average  use  of  time,  the  gaps  in  time  use  between  the   best-­‐  and  worst-­‐performing  teachers  in  each  school  is  much  smaller.  Top  teachers  in  these  Colombian   schools  spend  an  impressive  97  percent  of  total  class  time  on  instruction,  but  even  the  least  efficient   teachers  in  these  schools  achieve  63  percent  of  time  on  instruction:  a  34  percentage-­‐point  difference   that  is  roughly  half  the  size  of  the  within-­‐school  gap  in  the  bottom  quintile  of  schools.  What  truly   distinguishes  top  schools  is  consistency:  good  schools  have  less  variation  from  classroom  to  classroom  in   one  very  basic  parameter  of  teacher  performance:  the  share  of  class  time  used  for  instruction.  Whether   the  country  has  relatively  high  across-­‐school  variation  (such  as  Honduras,  Jamaica,  or  Brazil)  or  low   (Mexico  and  Colombia)  the  top  quintile  of  the  distribution  is  characterized  by  more  consistent  teacher   performance  across  different  classrooms.     The  implications  of  these  data  are  profound.  First,  school  systems  clearly  are  not  focused  on  the  issue  of   instructional  time.  In  some  sense,  this  is  understandable.  No  LAC  school  system  today  collects   standardized  data  on  classroom  dynamics;  it  is  understandable  that  large  differences  can  persist  in  the   absence  of  any  detection  system.     But  variations  in  classroom  practice  within  a  school  are  another  matter.  Direct  observation  of  all  the   classrooms  within  a  single  school  is  not  only  technically  feasible  but  is  an  implicit  responsibility  of  school   directors.  There  is  clearly  large  scope  for  directors  to  promote  more  exchange  of  practice  within  their   schools.  The  costs  of  identifying  the  most  effective  teachers  within  a  school  and  ensuring  that  other   teachers  observe  and  learn  from  these  examples  are  tiny  compared  to  the  costs  of  traditional  teacher   training  programs,  which  require  bringing  large  numbers  of  teachers  off-­‐site  and  hiring  trainers.     Raising  the  average  quality  teacher  practice  across  schools  is  the  responsibility  of  system  managers.   Many  different  approaches  can  be  imagined:  feeding  back  comparative  classroom  observation  data  to   schools  as  an  input  to  their  development  planning;  initiating  new  forms  of  teacher  training  based  on   videotaped  examples  of  good  and  weak  teacher  practice;  including  assessments  of  teachers’  classroom   practice—whether  by  video  or  trained  observers—in  teacher  performance  evaluations.  Policymakers  in   the  seven  countries  involved  in  this  program  have  taken  the  results  as  a  stimulus  to  action  along  several   of  these  lines.  Our  research  created  a  baseline  picture  of  what  LAC  students  encounter  inside  the   classroom  today.  It  also  provides  a  basis  for  tracking  how  ongoing  and  new  reforms  in  these  countries   succeed  in  reshaping  that  reality.     Recruiting  better  teachers   There  are  three  core  challenges  in  raising  teacher  quality:  recruiting,  grooming,  and  motivating  better   teachers.  Of  these,  recruiting—raising  the  caliber  of  teachers  at  the  point  of  recruitment—is  likely  to  be   the  most  complicated  for  LAC  countries,  because  it  depends  on  raising  the  selectivity  of  teaching  as  a   profession.  Global  research  on  high-­‐performing  education  systems  consistently  points  to  the  ability  to   attract  top  talent  into  teaching  as  a  critical  underlying  factor  that  takes  education  systems  from  “good  to   great”  (Barber  and  Mourshed  2007).  But  attracting  highly  talented  individuals  into  teaching  requires   aligning  a  complex  and  interrelated  set  of  factors  that  can  be  difficult  and  slow  to  change,  including   salaries  and  the  salary  structure,  the  prestige  of  the  profession,  the  selectivity  of  entry  into  teacher   education,  and  the  quality  of  that  education.     Raising  selectivity.  Our  analysis  suggests  that  salary  increases  may  be  necessary  in  some  countries  to   make  or  keep  teachers’  average  salaries—and  the  structure  of  teachers’  salary  incentives—competitive.   But  salary  increases  will  raise  quality  only  if  they  are  accompanied  by  policies  to  raise  the  selectivity  of   teacher  education  programs.  This  is  a  crucial  issue  in  the  LAC  region  that  gets  far  too  little  attention.  In   Singapore  and  Finland,  only  20  percent  of  secondary  school  students  who  apply  to  teacher  education       programs  are  accepted,  and  all  come  from  the  top  third  of  students.  In  LAC,  there  is  virtually  no   winnowing  of  teacher  candidates  at  the  point  of  entry  into  teacher  education,  and  academic  standards   are  lower  than  in  other  professional  fields.  As  a  result,  the  share  of  tertiary  education  students  in  Latin   America  enrolled  in  teacher  training  is  much  higher  than  in  many  OECD  countries,  and  many  LAC   countries  are  currently  producing  an  excess  of  teacher  graduates  (figure  O.14).  In  Peru,  only  50  percent   of  teacher  graduates  find  jobs  as  classroom  teachers;  in  Costa  Rica,  only  54  percent.  Chile’s  Ministry  of   Education  in  2013  estimated  that  as  many  as  half  of  all  students  graduated  from  teacher  education   programs  in  recent  years  are  currently  employed  in  retail.     Figure  O.14  Tertiary  graduates  who  studied  education   percent   45%   40%   35%   30%   25%   20%   15%   10%   5%   0%   Aruba   Bermuda   Costa   Salvador   Guyana   Mexico   Panama   Chile   Uruguay   Finland   South   Korea   Japan   Brazil   Argenln Barbado Hondura Colombi Rica   El   a   a   s   s     Source:  UNESCO  2012.  Data  is  for  most  recent  available  year  between  2009  and  2012.   Since  teacher  education  is  often  subsidized  by  the  public  sector,  producing  an  excess  of  teachers  diverts   resources  from  other,  more  productive  investments  in  student  learning.  In  countries  where  students   finance  teacher  education  with  their  own  resources  or  loans,  the  overproduction  of  graduates  from  a   four-­‐  or  five-­‐year  program  of  study  that  does  not  lead  to  relevant  employment  can  be  even  more   problematic,  and  may  generate  social  unrest.  Most  fundamentally,  a  lack  of  selectivity  undermines  the   prestige  of  the  profession  and  makes  teacher  education  less  attractive  for  top  students.     The  selectivity  and  status  of  the  teaching  profession  are  not  immutable  qualities  of  an  education  system.   Finland  made  raising  teacher  selectivity  a  cornerstone  of  an  education  reform  strategy  adopted  in  the   1970s.  Over  several  decades,  it  transformed  its  labor  market  for  teachers  from  one  where  a  large   number  of  teacher  training  institutions  of  variable  quality  produced  an  excessive  number  of  teachers  to   one  where  a  much  smaller  number  of  high-­‐quality  institutions  produce  just  enough  high-­‐talent  teachers,   all  of  whom  find  teaching  positions  and  enjoy  high  social  prestige.     Global  experience  points  to  three  key  levers  for  making  teacher  recruitment  more  selective:     • Raising  standards  for  entry  into  teacher  education   • Raising  the  quality  of  teacher  education  schools   • Raising  hiring  standards  for  new  teachers   Raising  entry  standards  for  teacher  education.  The  principle  of  university  autonomy  in  Latin  America   legally  prevents  most  ministries  of  education  from  directly  controlling  admissions  standards  for  pre-­‐ service  teacher  training.  A  few  prestigious  universities  are  selective,  but  in  most  countries  the  majority   of  new  teachers  are  produced  by  low-­‐quality  private  providers  and  nonuniversity  teacher  training   institutes  subject  to  weak  quality  assurance.         Four  main  strategies  are  being  pursued  by  education  ministries  in  the  region  to  address  these  issues:  (a)   closing  low-­‐quality  schools  under  direct  control  of  the  ministry  (typically  nonuniversity  teacher   education  institutions);  (b)  establishing  a  national  teacher  university  directly  controlled  by  the  ministry,   such  as  Singapore’s  National  Institute  for  Education;  (c)  creating  special  scholarships  for  top  students;   and  (d)  raising  accreditation  standards  for  autonomous  tertiary  institutions,  forcing  closure  or   adaptation.     Peru  tackled  the  oversupply  of  low-­‐quality  teachers  from  nonuniversity  teacher  training  schools   (Institutos  Superiores  Pedagogicos  [ISP])  in  2006  with  the  introduction  of  a  national  bar  for  admissions.   Requiring  candidates  to  achieve  a  minimum  competency  score  on  cognitive  tests,  a  writing  test,  and  an   interview  had  a  dramatic  effect  on  ISP  enrollments;  they  dropped  from  11,000  to  389  in  a  single  year   (figure  O.15).  A  number  of  regional  institutes  were  suddenly  threatened  with  closure,  raising  concerns   about  potential  teacher  shortages  in  bilingual  and  rural  communities.  In  2012,  the  ministry  returned   control  over  admissions  to  the  institutions,  but  with  annual  enrollment  caps  set  by  the  ministry.     Figure  O.15  Raising  the  bar  for  teacher  education  programs  in  Peru,  2006–10   Thousands  of  new  enrollments   16,000   14,000   12,000   Students   10,000   8,000   Non-­‐university   6,000   4,000   University   2,000   0   2006   2007   2008   2009   2010   Entrance  year     Source:  World  Bank  construction  from  Peru  Ministry  of  Education  data.   Ecuador  has  been  similarly  aggressive  in  trying  to  raise  the  quality  of  teacher  education.  It  closed  14  low-­‐ quality  teacher  preparation  institutions  in  2012  and  is  creating  a  high-­‐level  pedagogic  university   dependent  on  the  Ministry  of  Education,  the  Universidad  Nacional  de  Educación  (UNAE).  Researchers   have  identified  a  “tight  coupling”  between  the  Ministry  of  Education  and  the  institutions  where  teachers   are  educated  as  a  factor  in  the  educational  success  of  countries  as  different  as  Singapore  and  Cuba   (Carnoy  2007).  “Tight  coupling”  ensures  the  coordination  of  teacher  education  with  national  education   policy  goals,  such  as  higher  selectivity  at  entry  and  stronger  emphasis  on  math  instruction,  critical   thinking,  and  twenty-­‐first-­‐century  information  technology  skills.  The  UNAE  will  prepare  teachers  for  a   new  national  curriculum  currently  being  designed  in  consultation  with  national  and  international   stakeholders  and  experts.  Another  central  idea  is  that  UNAE  will  become  the  link  between  national   policymakers  and  other  teacher  training  institutions,  seeding  the  latter  over  time  with  highly  qualified   faculty  trained  at  the  UNAE.     A  third  strategy,  which  offers  shorter-­‐term  impact,  is  the  use  of  targeted  incentives  to  attract  top   secondary  school  graduates  into  teaching.  In  Singapore,  Finland,  Hong  Kong,  and  Scotland,  teacher   training  is  open  only  to  select  candidates,  but  these  students  receive  free  tuition  plus  a  salary  or  stipend   while  they  are  in  training  (Garland  2008).  Colombia  and  Chile  have  recently  launched  similar  programs.   Chile’s  program,  called  the  Beca  Vocación  de  Profesor  (BVP),  in  2010  began  offering  full  tuition  for   students  who  score  600  or  higher  on  the  university  entrance  examination  (PSU),  and  agree  to  study       education  and  work  as  a  full-­‐time  teacher  for  at  least  three  years  in  public  or  subsidized  schools  after   graduating.  About  3,500  students  per  year  have  qualified  for  the  Beca  since  2010,  a  relatively  small   share  of  the  130,000  students  in  teacher  education  in  Chile.  But  researchers  have  confirmed  the  BVP’s   success  in  attracting  students  with  a  stronger  academic  profile.  Feedback  from  top  universities  is  that   the  new  students  are  stimulating  higher  academic  performance  from  their  classmates  as  well.     The  fourth  and  most  important  tool  is  the  accreditation  and  review  process.  Developing  the  institutional   capacity  for  a  national  higher  education  accreditation  system  is  complex  and  takes  time.  Chile’s   experience,  however,  shows  that  accreditation  information  can  exert  important  influence  on  student   enrollment  choices  (figure  O.16).  When  mandatory  accreditation  reviews  of  teacher  education  programs   began  in  2006,  80  percent  of  the  country’s  940  teaching  programs  received  either  no  accreditation  or   the  lowest  category.  Even  without  direct  action  to  close  low-­‐quality  institutions,  in  the  space  of  a  few   years,  the  teacher  training  market  shifted  massively  from  77  percent  of  enrollments  in  nonaccredited   programs  to  70  percent  of  enrollments  in  accredited  programs.  Proposed  legislation  goes  further,   requiring  that  all  teachers  hired  into  public  (or  publicly  subsidized)  schools  must  have  graduated  from   accredited  programs.     Figure  O.16  Impact  of  accreditation  information  on  teacher  education  enrollments  in  Chile,  2007–10     (full-­‐time  students  enrolled,  thousands)   100000   80000   60000   40000   20000   0   2007   2008   2009   2010   Accredited  programs   Non-­‐accredited  programs     Source:  World  Bank  construction  using  Chile  Ministry  of  Education  enrollment  data.     Raising  the  quality  of  teacher  education.  The  academic  quality  of  students  entering  teacher  training  is   weak,  but  the  quality  of  those  programs  is  also  dismally  low.  Qualitative  accounts  of  pre-­‐service  training   in  Latin  America  generally  describe  it  as  failing  to  provide  sufficient  content  mastery  and  student-­‐ centered  pedagogy;  being  isolated  from  the  school  system  and  education  policy  making;  and  including   practical  exposure  to  work  in  schools  only  toward  the  end  of  the  degree  and  sometimes  not  at  all   (UNESCO  2012).   While  global  research  suggests  that  pre-­‐service  training  programs  focused  on  the  work  teachers  will   actually  face  in  classrooms  lead  to  more  effective  first-­‐year  teachers  and  higher  learning  for  their   students  (Boyd  et  al.  2009),  most  Latin  American  and  Caribbean  countries  do  not  set  a  minimum   standard  for  practice  teaching  and  leave  it  to  institutions  to  define.  As  a  result,  while  in  high-­‐performing   Cuba  72  percent  of  a  teacher  education  program  is  spent  doing  practice  in  schools  (i.e.,  more  than  7,100   hours  over  5  years),  national  thresholds  in  other  countries  require  far  less.  (figure  O.17).                     Figure  O.17  Compulsory  pre-­‐service  teaching  practice  in  LAC  countries     (number  of  weeks)   160   140   120   100   80   60   40   20   0   Cuba   Mexico   Argenlna  U.   Argenlna   Brazil   Peru  N.U.   N.U.     Source:  Franco  2012.     Note:  N.U.refers  to  non-­‐university  programs.    U.  refers  to  university  level  programs.  Figures  assume  40  program  hours   per  week  and  40  weeks  per  year.   University  autonomy  prevents  directly  mandating  changes  in  the  content  of  teacher  education,  but   several  ministries  of  education  in  the  region  are  creatively  using  competitive  funding  programs  to   stimulate  such  reforms.  Chile  in  2013  launched  a  new  line  of  competitive  funding  to  support  a  “rethink”   of  teacher  education.  Its  ministry  is  open  to  proposals  for  major  change,  including  shortening  the   number  of  years  required,  radically  changing  curriculum  content,  and  increasing  the  time  spent  working   in  classrooms.  The  only  requirement  is  that  proposals  be  grounded  in  global  research  evidence.  Peru’s   competitive  fund  for  tertiary  education  (Fondo  de  Estímulo  de  la  Calidad)  also  has  a  specific  line  of   support  to  improve  the  quality  of  teacher  training  institutions.     Raising  hiring  standards.  The  low  quality  of  teacher  education  programs  makes  it  important  that  public   education  systems  screen  effectively  at  the  point  of  hiring.  Three  main  policy  instruments  can  ensure   this:  (a)  national  teacher  standards;  (b)  preemployment  tests  of  teachers’  skills  and  competencies;  and   (c)  alternative  certification.   National  teacher  standards,  articulating  “what  a  teacher  should  know  and  be  able  to  do,”  are  an   important  step  in  the  development  of  a  more  professional  teaching  corps.  Over  the  past  twenty  years,   most  OECD  countries  have  put  serious  effort  into  developing  standards  for  teachers.  Relatively  few  LAC   countries  have  done  so.     The  exception  is  Chile,  which  adopted  national  teacher  standards  (Marco  para  la  Buena  Enseñanza)  in   2003,  after  three  years  of  joint  work  by  a  national  commission  and  the  teachers’  union.  Chile’s   framework  remains  a  best-­‐practice  example  for  the  region  (table  O.3)  It  has  guided  the  design  of  other   key  policies,  including  Chile’s  teacher  evaluation  system  and  teacher  exit  exam,  the  Prueba  Inicia.   National  teaching  standards  take  time  and  effort  to  develop,  but  formally  establishing  high  standards  for   teachers  lays  the  foundation  for  education  quality.               Figure  O.3  Chile’s  Framework  for  Good  Teaching   Preparation   Learning  environment   -­‐  Knowledge  of  the  content  and  curriculum   -­‐  Creates  an  environment  of  acceptance,   -­‐  Knows  the  students   equity,  con`idence,  and  respect   -­‐  Dominates  the  didactics   -­‐  Manifests  high  expectations  about  all   students'  learning  and  development   -­‐  Organizes  objectives  and  contents   possibilities.     coherently   -­‐  Classroom  norms     -­‐  Coherent  evaluation  strategies   -­‐  Organized  work  environment  and  physical     setting   Professional  responsibilities   Instruction  for  all  students   -­‐  Re`lects  systematically  on  teaching   -­‐  Communicates  learning  objectives   -­‐  Professional  and  team  relationships  with   -­‐  Uses  challenging,  coherent,  and  signi`icant   colleagues   teaching  strategies   -­‐  Takes  on  responsibility  in  orienting   -­‐  Content  of  class  is  treated  with  conceptual   students   thoroughness  and  is  comprehensible  for  all   -­‐  Favors  collaboration  with  parents  and   students   guardians   -­‐  Optimizes  use  of  time   -­‐  Knows  updated  information  about  the   -­‐  Promotes  thinking   profession,  the  education  system,  and   current  policies   -­‐  Evaluates  and  monitors  student  learning     Source:  Adapted  from  Chile  Ministry  of  Education  2008.   A  second  tool  is  certification  exams  or  competency  tests  to  screen  teacher  candidates.  No  country  in   Latin  America  and  the  Caribbean  today  has  a  compulsory  certification  process  for  the  teaching   profession  as  stringent  as  those  used  in  medicine,  law,  or  accounting—i.e.,  a  rigorous  examination  that   prohibits  those  who  do  not  meet  its  standards  from  being  hired  anywhere  into  the  profession  (public  or   private  sector).  But  Colombia  and  El  Salvador  have  introduced  mandatory  exit  exams  which  teacher   graduates  must  pass  in  order  to  be  hired  by  the  public  sector.    In  El  Salvador  there  has  been  clear   progress  in  raising  the  share  of  teacher  graduates  who  pass  the  exam  (Evaluación  de  las  Competencias   Académicas  y  Pedagógicas  [ECAP])  since  it  was  introduced  in  2001  (figure  O.18).  An  important  further   step  will  be  research  on  how  the  flow  of  better-­‐prepared  teachers  into  the  school  system  affects  student   learning  results.     In  Colombia,  the  exam  for  graduating  teachers  established  under  the  2002  reform  law  also  set  a  higher   bar.    But  the  share  of  teachers  hired  under  the  new  system  remains  relatively  small,  and  there  is   unfortunately  little  evidence  on  their  effectiveness.                     Figure  O.18  Teacher  exit  exam  pass  rate  in  El  Salvador,  2001–12   percent  of  new  teacher  graduates  with  passing  score   100%   90%   80%   70%   60%   50%   40%   30%   20%   10%   0%   2001   2002   2003   2004   2005   2006   2007   2008   2009   2010   2011   2012     Source:  Picardo  Joao  2012.   Introducing  mandatory  certification  exams  that  set  a  higher  bar  for  teacher  quality  typically  requires  a   transitional  period.  When  New  York  State  established  a  more  stringent  teacher  certification  exam  in   1998,  a  high  share  of  new  teacher  graduates  could  not  pass  it.  The  state  permitted  a  five-­‐year  window   for  school  districts  to  hire  “temporary  license”  teachers  who  had  not  passed  the  exam.  The  new   standards  hit  hardest  in  high-­‐poverty  urban  school  districts  such  as  New  York  City,  where  temporary   license  teachers  accounted  for  63  percent  of  new  hires  in  the  poorest  quartile  of  schools.  To  meet  the   2003  deadline  for  eliminating  such  hiring,  the  city  was  forced  to  pursue  aggressive  strategies  to  attract   higher  quality  teachers.  But  by  2005,  it  succeeded  in  eliminated  the  hiring  of  uncertified  teachers,   sharply  narrowed  the  gap  in  teacher  qualifications  between  high-­‐  and  low-­‐poverty  schools,  and  saw   large  improvements  in  student  performance  in  high-­‐poverty  schools  (Boyd  et  al.  2008).   While  mandatory  certification  exams  are  the  most  powerful  instrument  for  raising  teacher  standards,   nonbinding  “exit  exams”  upon  completion  of  teacher  education  programs  can  also  support  better  hiring   choices,  monitor  the  quality  of  graduates  over  time,  and  expose  quality  differences  across  teacher   training  schools.  Chile’s  2008  “Inicia”  test  and  the  national  teaching  exam  being  developed  in  Brazil   (Prova  Nacional  de  Concurso  para  o  Ingresso  na  Carreira  Docente)  are  examples.  Making  such   examinations  voluntary  has  a  downside,  however;  in  Chile,  only  40  percent  of  graduates,  on  average,   opt  to  take  the  Inicia  test,  and  of  these,  70  percent  fail  to  achieve  the  threshold  score.  Ministry  concerns   that  the  60  percent  of  teacher  graduates  who  refuse  to  take  the  exam  are  even  less  prepared  to  meet   national  teacher  standards  sparked  a  legislative  proposal  to  make  the  exam  mandatory.     A  third  strategy  for  raising  the  quality  of  new  teachers  is  to  bypass  low-­‐quality  education  schools  entirely   and  recruit  teachers  trained  in  other  disciplines,  a  practice  known  as  alternative  certification.  This  was   key  to  New  York  City’s  rapid  progress  in  raising  teacher  quality  and  is  widely  used  in  other  U.S.  urban   school  districts  that  have  difficulty  attracting  teachers  to  work  with  disadvantaged  populations.     Rigorous  U.S.  studies  have  generally  concluded  that  students  of  teachers  with  alternative  certification,   most  notably  those  from  the  Teach  For  America  program,  do  as  well  or  better  than  students  of  regularly   recruited  teachers.     Most  countries  in  LAC  have  not  pursued  alternative  certification  on  any  scale,  although  it  is  permitted  in   Colombia,  and  proposed  in  Chile  and  Mexico.  Since  2007,  however,  several  LAC  countries  have  launched       national  branches  of  Teach  For  All,  modeled  on  Teach  for  America.  In  Chile,  Peru,  Mexico,  Colombia,   Argentina,  and  Brazil,  Teach  For  All  programs  have  recruited  top  university  graduates  from  other   disciplines  willing  to  commit  to  two  years  of  teaching  in  highly  disadvantaged  schools.     A  randomized  evaluation  of  the  program  in  Chile,  Enseña  Chile,  is  currently  underway,  with  first  results   expected  in  2016.  A  2010  study  documented  that  the  programs  in  Argentina,  Peru,  and  Chile  have   succeeded  in  attracting  high-­‐talent  university  graduates,  and  that  students  of  Enseña  Chile  teachers  had   higher  learning  outcomes  than  those  of  comparable  traditional  teachers.  Students  of  Enseña  teachers   also  had  better  socioemotional  competencies,  including  higher  self-­‐esteem  and  self-­‐efficacy.  The  Enseña   teachers  had  more  positive  attitudes  about  their  students’  ability  to  learn  and  higher  expectations  for   them  (Alfonso,  Santiago,  and  Bassi  2010).  Although  the  LAC  programs  currently  operate  on  a  small  scale,   they  appear  to  be  a  useful  tool,  especially  for  raising  teacher  quality  in  disadvantaged  urban  or  rural   schools  and  for  hard-­‐to-­‐fill  disciplines,  such  as  secondary  school  math  and  science.     Raising  teacher  selectivity  over  the  next  decade.  All  LAC  countries  face  the  challenge  of  recruiting   better  teachers,  but  it  will  play  out  in  different  countries  against  very  different  demographic  backdrops.   Over  half  of  the  region—including  all  of  its  largest  countries—will  see  student  population  fall  by  as  much   as  31  percent,  while  other  countries,  especially  in  Central  America  and  the  Dominican  Republic,  will  face   continued  growth  in  student  numbers.  UNESCO  (the  United  Nations  Educational,  Scientific,  and  Cultural   Organization)  projects  that  with  no  change  from  2010  in  enrollment  ratios  or  pupil-­‐teacher  ratios,  the   region  would  need  8  percent  fewer  teachers  by  2025.     Since  not  all  countries  in  the  region  have  achieved  universal  schooling  coverage,  especially  at  the   secondary  level  and  preschool,  the  demand  for  teachers  also  depends  on  assumptions  about  how   quickly  these  countries  expand  coverage.  To  project  a  scenario  of  maximum  potential  demand  for  new   teachers,  we  assumed  that  all  countries  in  the  region  reach  100  percent  primary  education  enrollment,   90  percent  secondary  enrollment,  and  90  percent  preschool  enrollment  of  children  ages  four  to  six  by   the  year  2025,  even  if  this  implies  rates  of  schooling  expansion  far  above  countries’  past  trends.     Under  this  highly  ambitious  scenario  of  schooling  expansion,  the  region  would  still  need  a  smaller  overall   stock  of  preschool,  primary,  and  secondary  teachers  in  2025,  assuming  that  current  pupil-­‐teacher  ratios   (PTR)  in  every  country  remain  stable.  The  total  stock  of  teachers  would  fall  from  7.35  million  in  2010  to   about  6.61  million  in  2025.  While  some  countries  would  need  to  increase  teacher  numbers  to  support   the  expansion  of  coverage,  others  would  see  only  modest  changes  or  large  declines.  A  net  decline  in   teacher  numbers  implies  an  opportunity  for  countries  to  pay  a  smaller  stock  of  teachers  higher  average   salaries,  which  could  help  raise  the  attractiveness  of  the  profession.  But  the  salary  increment  would  be   relatively  small,  and  even  this  potential  “fiscal  space”  would  materialize  only  if  current  pupil-­‐teacher   ratios  did  not  decrease.     However,  the  tendency  for  school  systems  with  a  declining  student  population  is  to  let  the  pupil-­‐teacher   ratio  decline.  It  requires  active  management  to  reduce  the  teaching  force  pari  passu  with  demographic   decline,  thus  maintaining  a  stable  PTR.  Both  teachers’  unions  (which  wish  to  protect  job  stability)  and   parents  (who  believe  that  a  smaller  class  size  is  better  for  their  children)  resist  this.  The  falling  pupil-­‐ teacher  ratios  already  observed  in  LAC  countries  with  declining  student  populations  indicate  that  this   pattern  has  already  taken  hold.     The  contrast  with  East  Asian  countries  is  sharp.  Singapore,  Korea,  China,  and  Japan  consciously  maintain   relatively  high  pupil-­‐teacher  ratios  to  free  up  resources  for  higher  teacher  salaries,  a  longer  school  day,   and  cost-­‐effective  nonsalary  investments.  Teacher  salaries  in  these  countries  are  relatively  high  on   average  and  are  differentiated  by  competency  and  performance,  which  attracts  more  talented   individuals.       To  explore  the  implications  of  a  similar  trade-­‐off  in  LAC,  we  projected  the  same  ambitious  trends  in   enrollment,  but  with  active  management  of  pupil-­‐teacher  ratios,  to  reach  target  levels  of  18  to  1  in   preschool  education  and  20  to  1  for  primary  and  secondary  education  by  2025.  While  some  countries   would  need  to  hire  more  teachers  to  meet  the  projected  goals  of  close  to  universal  coverage  and  lower   pupil-­‐teacher  ratios  than  they  currently  enjoy,  the  overall  size  of  the  teaching  force  in  LAC  would  decline   by  11  percent  (figure  O.19).     Figure  O.19  Change  in  stock  of  teachers  needed,  assuming  expanded  coverage  and  efficient  pupil-­‐ teacher  ratios,  2010–25   percent   70%   60%   50%   40%   30%   20%   10%   0%   -­‐10%   -­‐20%   -­‐30%   -­‐40%   -­‐50%   -­‐60%   -­‐70%   Cuba   Nicaragua   Ecuador   LAC   Chile   El  Salvador   Honduras   Dominican  Republic   Guatemala   Mexico   Paraguay   Brazil   Uruguay   Colombia   Argentina   Peru   Panama   Costa  Rica     Source:  World  Bank  projections,  using  data  from  UNESCO  2009  and  UNESCO  2011.   Note:  Projections  assume  that  all  countries  reach  target  GERs  (gross  enrollment  ratios)  of  100  percent  for  primary  and   90  percent  for  preschool  and  secondary  education  by  2025,  and  pupil-­‐teacher  ratios  of  18  to  1  for  preschool   education  and  20  to  1  for  primary  and  secondary  education  by  2025.   For  several  of  the  larger  countries  in  the  region,  policies  to  manage  the  pupil-­‐teacher  ratio  combined   with  demographic  trends  would  permit  significant  increases  in  teacher  salaries.  In  Brazil,  for  example,   this  scenario  results  in  a  27  percent  decline  in  the  number  of  teachers,  from  2.9  million  to  2.1  million,  by   2025.  But  it  would  allow  for  a  36  percent  real  increase  in  average  teacher  salaries  and  move  teachers’   relative  salaries  from  the  76th  percentile  of  the  wage  distribution  to  the  85th  percentile,  compared  to   the  90th  percentile  for  other  professional  workers.     In  the  coming  decade,  several  countries  will  have  a  unique  opportunity  to  raise  teacher  quality  through   higher  salaries,  stronger  incentives,  and  higher  nonsalary  spending.  There  would  be  no  increase  needed   in  overall  education  budgets  if  school  systems  carefully  manage  teacher  numbers  in  favor  of  teacher   quality.  Since  these  assumptions  are  based  on  constant  real  spending  per  student,  countries  that  have   declining  student  populations  and  raise  education  spending  as  a  share  of  GDP  would  have  even  more   resources  per  teacher  to  finance  a  move  to  higher  quality.   This  heterogeneity  will  create  diverse  challenges  for  teacher  policy.  Projected  declines  in  the  student   population  in  half  of  the  region—including  its  largest  countries—will  make  it  easier  for  school  systems  to       finance  higher  teacher  quality,  but  will  pose  the  political  challenge  of  pruning  low  performers  out  of  the   force  to  make  room  for  higher  quality  new  recruits.  With  declines  in  some  cases  of  20  percent  or  more   in  the  size  of  the  teaching  force  by  2025,  managing  both  teacher  exits  and  selective  recruitment  with  a   strategic  focus  on  quality  is  critical.  In  countries  where  increased  need  for  teachers  is  projected  over  the   next  decade,  the  major  challenge  is  the  financial  burden  of  raising  education  spending  to  support  the   recruitment  of  new  teachers  at  higher  standards.  For  these  countries,  efficient  class  size  is  a  critical   policy  choice.   Grooming  Great  Teachers   Once  teachers  have  been  hired,  it  is  the  task  of  a  school  system  to  make  them  as  effective  as  possible.   This  involves  assessing,  managing,  and  supporting  individual  teachers’  development  of  their  craft  and   building  a  professional  community  of  teachers,  both  within  schools  and  across  the  school  system.  Four   essential  tasks  are:   • Induction:  support  for  teachers’  development  during  their  critical  first  five  years  of  teaching     • Evaluation:  systems  for  regular  assessment  of  individual  teachers’  strengths  and  weaknesses   • Professional  development:  effective  training  to  remedy  teachers’  identified  weaknesses  and   leverage  the  skills  of  top  performers     • Management:  matching  teacher  assignments  to  schools’  and  students’  needs,  and  building   effective  schools  through  shared  practice  and  professional  interaction     Teacher  induction.  A  consistent  finding  of  education  research  is  that  new  teachers  face  a  steep  learning   curve  in  their  first  three  to  five  years  on  the  job  (Boyd  et  al.  2009;  Chingos  and  Peterson  2010;  Hanushek   and  Rivkin  2010).  During  this  window,  school  systems  have  an  important  opportunity  to  support  and   maximize  the  development  of  new  teachers  and  to  identify  those  who  should  be  counseled  out  of  the   profession.  Because  teachers  are  typically  hired  into  civil  service  positions,  which  make  dismissal  on   performance  grounds  difficult  once  they  are  confirmed,  there  is  a  high  payoff  to  avoiding  recruitment   mistakes.  Both  of  these  goals  are  served  by  a  well-­‐organized  induction  program  and  the  effective  use  of   probationary  periods.   In  the  LAC  region,  very  few  countries  outside  of  the  English-­‐speaking  Caribbean  countries  have  formal   induction  programs  for  entering  teachers.  Belize’s  program,  which  over  the  course  of  a  teacher’s  first   year  includes  tutoring,  observation  in  the  classroom,  mentoring  support,  action-­‐research  projects,  and   assessments,  has  a  particularly  impressive  design.   Effective  induction  goes  hand  in  hand  with  consequential  probationary  periods.  Most  OECD  countries   use  probationary  periods.  Some  school  systems  in  the  United  States  have  extended  these  to  three  years   or  more  to  allow  more  time  to  assess  teachers’  performance  and  growth  potential  before  making  a  final   contract  decision.  But  in  Latin  America  and  the  Caribbean  probationary  periods  with  consequences  are   rare.  Only  two  of  seven  countries  reviewed  recently  (Colombia  and  the  Dominican  Republic)  have   consequential  probationary  periods  that  include  a  comprehensive  assessment  of  new  teachers  (Vaillant   and  Rossel  2006).  A  2013  program  introduced  in  Rio  de  Janeiro  municipality  presents  a  promising  model:   new  teacher  candidates  are  given  an  intensive  training  course  in  effective  classroom  dynamics—which   draws  on  the  municipality’s  Stallings  results—and  are  subsequently  observed  teaching  and  evaluated   before  their  recruitment  is  confirmed.       Teacher  evaluation.  Top  education  systems  invest  heavily  in  the  evaluation  of  teacher  performance.   Teacher  evaluation  plays  two  critical  roles:  improving  teachers’  quality  and  holding  them  accountable.   Singapore,  Japan,  Korea,  and  China’s  Shanghai  all  have  effective  systems  for  assessing  their  teachers’   performance  and  progress.    The  experience  in  Latin  America  to  date  is  much  more  limited.  While  Mexico       (in  the  1990s)  and  Colombia  in  2002  introduced  the  region’s  first  teacher  evaluation  systems   implementation  issues  have  undermined  their  impact  and  Mexico’s  system  is  currently  being   redesigned.  (Vegas  et  al.  2005,  Ome  2012)  Chile’s  2003  system  remains  the  region’s  best  practice   example  to  date.  Ecuador  began  implementing  teacher  performance  evaluations  in  2007,  and  Peru  is   currently  designing  a  comprehensive  system  similar  to  Chile’s.  Elsewhere  in  the  region,  countries  have   some  elements  in  place,  but  they  are  less  comprehensive  and  systematic.   Putting  in  place  a  sound  system  of  teacher  evaluation  is  expensive  and  institutionally  challenging,  but  it   can  make  other  important  education  system  functions  more  efficient.  Teacher  performance  information   can  make  in-­‐service  training  investments  more  relevant,  help  target  them  better,  and  even  reduce   overall  training  costs.  It  creates  the  information  base  needed  for  individual  performance  incentives  and   accountability  measures.  It  provides  teachers  with  individualized  feedback  that  might  not  be  as   forthcoming  or  insightful  from  their  direct  supervisors  and  peers,  and  increases  teachers’  motivation  to   pursue  training  or  professional  improvement.  If  these  benefits  are  captured,  the  net  costs  of  a  good   teacher  evaluation  system  may  be  small.     Growing  global  experience  with  teacher  evaluation  points  to  four  key  features  of  successful  systems.   First,  they  are  grounded  in  teacher  standards:  a  clear  articulation  of  the  competencies  and  behaviors   that  good  teachers  are  expected  to  have  and  demonstrate.  Second,  they  measure  performance   comprehensively.  A  three-­‐year  research  program  in  the  United  States  concluded  that  a  reliable   judgment  about  an  individual  teacher’s  performance  requires  multiple  measures,  including  classroom   observation  (ideally  multiple  times)  and  student  and  peer  feedback.  In  combination,  these  measures  can   produce  assessments  of  teachers’  effectiveness  that  are  well-­‐correlated  with  their  students’  value-­‐ added  learning  gains  (Kane  et  al.  2012).  Chile’s  teacher  evaluation  system  is  a  good  example  of   comprehensive  evaluation.  It  combines  an  observation  of  teachers’  classroom  practice  (through  a   videotaped  class),  a  sample  lesson  plan,  a  self-­‐assessment,  a  peer  interview,  and  an  assessment  by  the   school  director  and  pedagogical  supervisor.     Third,  good  evaluation  systems  use  instruments  that  have  technical  validity  and  protect  the  integrity  of   evaluation  processes.  The  implementation  of  Chile’s  system  is  contracted  through  competitive  bidding   to  an  independent  education  research  group  that  conducts  ongoing  research  to  improve  the  robustness   of  the  system.  The  evaluation  team  is  responsible  for  assuring  the  quality  and  consistency  of   implementation.  For  example,  it  provides  extensive  training  to  the  supervisors  who  evaluate  the  videos   of  teachers’  classroom  practice  and  the  teachers  who  conduct  peer  interviews.     Fourth,  good  systems  ensure  that  evaluation  results  have  consequences  for  teachers:  both  positive  and   negative.  Most  OECD  countries  use  their  evaluation  systems  as  a  platform  for  performance  incentives:   identifying  and  rewarding  top  performers.  In  Chile  and  Ecuador,    teachers  who  receive  outstanding   evaluations  are  eligible  for  bonus  pay.  Evaluation  systems  also  provide  the  soundest  basis  for  long-­‐term   grooming  of  individual  teachers’  potential  and  the  fairest  basis  for  promotions.  Rather  than  promoting   teachers  on  the  basis  of  seniority  alone—as  most  LAC  countries  currently  do—teachers  can  be   promoted  on  the  basis  of  recognized  competence.  A  salary  structure  aligned  with  evaluated   performance  creates  the  right  incentives  for  current  teachers  and  makes  the  profession  more  attractive   to  talented  candidates  in  the  future.       Effective  teacher  evaluation  systems  strengthen  accountability.  They  allow  school  system  managers  to   identify  teachers  in  need  of  improvement  and  create  strong  incentives  for  these  teachers  to  pursue  the   training  offered  and  apply  it  to  their  work.  In  Chile,  Ecuador  and  Colombia,  and  under  new  proposals  in   Peru  and  Mexico,  teachers  who  receive  poor  evaluations  are  offered  training  and  required  to  be   reevaluated.         Finally,  teacher  evaluation  gives  school  systems  data  to  deal  with  consistent  poor  performers   forthrightly  and  transparently.  In  Chile,  Ecuador  and  Colombia,  and  under  new  proposals  in  Peru  and   Mexico,  teachers  who  receive  successive  evaluations  in  the  lowest  performance  categories  are   dismissed  from  service.  An  education  system’s  capacity  to  identify  its  least-­‐effective  teachers  is  a   powerful  tool  for  raising  schooling  quality.  Research  suggests  that  systematically  targeting  the  lowest  5   percent  of  teachers  annually  for  “de-­‐selection”  can  produce  large  gains  in  student  learning  over  time   (Hanushek  2011,  Chetty  et  al.  2014).     A  rigorous  evaluation  of  the  impacts  of  the  Washington,  DC,  teacher  evaluation  system,  which  is   considered  a  best-­‐practice  model  in  the  United  States,  concluded  that  the  system  has  produced  an   impressive  improvement  in  overall  teacher  quality  in  just  the  first  three  years  of  implementation.   Researchers  documented  four  main  mechanisms:  voluntary  attrition  of  teachers  with  low  performance   ratings  increased  by  over  50  percent;  low-­‐rated  teachers  who  decided  to  stay  in  the  system  achieved  big   improvements  in  performance;  a  higher  share  of  top  teachers  stayed  in  the  system  (rather  than   transferring  to  other  school  districts);  and  teachers  at  the  threshold  for  bonuses  made  big   improvements  (Dee  and  Wyckoff  2013).  Since  the  teacher  evaluation  system  was  introduced,  not  only   teacher  quality  has  improved;  student  learning  gains  have  been  the  largest  of  any  urban  district  in  the   United  States.   Teacher  professional  development.  When  the  costs  of  teacher  time  are  included,  in-­‐service  training  is  a   major  element  of  education  spending  in  LAC.  In  Brazil  and  Mexico,  many  teachers  participate  in  more   than  one  month  of  training  annually.  Yet  evidence  on  the  cost-­‐effectiveness  of  training  is  almost   nonexistent.  The  global  evidence  base  is  also  limited.  The  most  common  conclusion  of  metastudies  is   that  the  relevance  of  training  content,  the  intensity  and  duration  of  the  course,  and  the  quality  of  the   delivery  are  key:  observations  that  abstract  from  the  central  question  of  how  to  design  relevant  program   content.     A  review  of  the  academic  literature  and  different  training  approaches  that  figure  prominently  in  the   “improving”  education  systems  identified  by  Mourshed,  Chijoke,  and  Barber  (2010)  suggests  that  four   broad  strategies  for  teacher  training  are  most  relevant  for  LAC  countries:     • Scripted  approaches:  training  to  prepare  teachers  in  low-­‐capacity  environments  to  use  specific   teaching  strategies  and  accompanying  materials  in  the  delivery  of  a  well-­‐defined  daily  curriculum   • Content  mastery:  training  focused  on  filling  gaps  or  deepening  teachers’  expertise  in  the  subjects   they  teach  and  how  to  teach  them  effectively   • Classroom  management:  training  focused  on  improving  teachers’  classroom  effectiveness  through   lesson  planning,  efficient  use  of  class  time,  strategies  for  keeping  students  engaged,  and  more   effective  teaching  techniques   • Peer  collaboration:  school-­‐based  or  cross-­‐school  opportunities  for  small  groups  of  teachers  to   observe  and  learn  from  each  other’s  practice  and  collaborate  on  curriculum  development,  student   assessment  strategies,  research,  and  other  activities  that  contribute  to  system  quality  and   teachers’  professional  development   Scripted  training  programs  are  relevant  for  many  LAC  countries,  and  perhaps  especially  for  early  grade   literacy  instruction  and  math  skills.  Honduras’  SAT  (Sistema  de  Aprendizaje  Tutorial)  scripted  training   program  for  middle  school  teachers  in  rural  areas  has  produced  higher  learning  results  at  lower  per-­‐ student  costs  (McEwan  forthcoming).  Scripted  approaches  have  also  been  used  successfully  in   Colombia’s  Escuela  Nueva  to  support  teachers  in  multigrade  schools,  and  in  the  Brazilian  states  of  Ceara   and  Minas  Gerais  to  train  early  grade  reading  teachers.  By  providing  teachers  with  comprehensive   support  in  the  use  of  teacher  guides,  lesson  plans,  classroom  reading  books,  and  reading  assessments  to       be  applied  at  regular  intervals,  Ceara    has  seen  significant  improvements  in  both    reading  and  math   results  (Costa  and  Carnoy  2014).  The  federal  Ministry  of  Education  is  now  supporting  national  scale-­‐up   of  this  approach.     Given  the  weak  content  mastery  of  many  teachers  in  LAC,  training  in  this  area  is  also  clearly  relevant.   Unfortunately,  there  is  no  rigorous  evaluation  evidence  of  successful  LAC  programs..  Useful  training  in   the  US  context,  however,  has  focused  on  the  specific  math  content  knowledge  required  for  effective   teaching  at  different  levels.  (Thames  and  Ball  2010)   Improving  teachers’  classroom  practice  emerged  as  a  clear  issue  from  the  classroom  observations   conducted  for  this  report,  including  teachers’  ability  to  use  class  time  and  materials  effectively  and  to   keep  students  engaged.  It  is  encouraging  that  a  number  of  LAC  school  systems  are  developing  courses   focused  on  classroom  management  techniques  that  can  improve  student  engagement  and  learning   performance.  In  several  cases,  governments  are  planning  rigorous  evaluations,  with  random  assignment   of  teachers  to  different  training  options  and  careful  measurement  of  both  subsequent  classroom   practice  (through  classroom  observations)  and  impacts  on  student  learning.  These  experiences  could   contribute  enormously  to  the  evidence  base  to  guide  effective  training  investments  not  only  in  LAC  but   also  globally.     Peer  collaboration—as  practiced  in  Finland,  Ontario  (Canada),  and  under  Japan’s  “Lesson  Study”  method     —is  integral  to  the  development  of  greater  professionalism  among  LAC’s  teachers  and  to  the  informal   exchange  of  practice  at  the  school  level  that  is  the  most  cost-­‐effective  strategy  for  improving  school   results.  There  is  an  incipient  trend  in  this  direction  in  several  LAC  countries.  For  example,  in  Ecuador   training  needs  are  identified  at  the  school  level,  and  colleagues  receive  training  together  (Ministerio  de   Educacion  del  Ecuador  2012).  Under  Peru’s  teacher  mentoring  program,  external  coaches  work  with  all   of  the  teachers  in  a  school  as  a  team,  providing  real-­‐time  feedback  and  advice  grounded  in  the  coaches’   observation  and  understanding  of  the  school’s  context  and  specific  challenges.  Under  Rio  municipality’s   new  Gente  and  Ginasio  carioca  experimental  programs,  the  school  day  has  been  extended  to  free  up   time  for  teacher  collaboration  and  team  teaching.  All  of  these  represent  very  new  approaches  for  the   LAC  region  and  have  yet  to  be  evaluated.  But  the  emphasis  placed  on  looking  inside  schools  and   classrooms  to  identify  the  issues  where  teachers  most  need  support  is  promising.   Designing  and  delivering  capacity-­‐building  programs  of  the  caliber  and  scale  required  in  most  countries   will  not  be  easy.  As  Carnoy  (2007)  has  observed,  “weak  coupling”  between  education  ministries  and   university  education  departments  makes  the  latter  ill-­‐prepared  to  respond  to  ministries’  needs.   Increasingly,  ministries  (and  secretariats  in  Brazil)  are  creating  their  own  in-­‐service  teacher  training   institutes  to  take  direct  control  of  the  content  and  delivery  of  teacher  professional  development.    While   it  is  too  early  to  evaluate  these  institutes’  effectiveness,  there  is  a  visible  shift  toward  professional   development  programs  that  directly  address  identified  issues.  Using  teacher  evaluation  data  to   determine  training  priorities;  exploiting  partnerships  with  NGOs,  think  tanks,  and  other  providers   operating  outside  of  university  education  departments;  and  investing  in  rigorous  evaluation  of  at  least   the  most  important  training  initiatives  will  help  make  investments  in  this  area  more  cost-­‐effective,   which  is  essential  for  faster  progress  in  raising  the  caliber  of  the  current  stock  of  teachers.     Teacher  deployment  and  management.  Grooming  teachers  so  they  develop  their  full  potential  and   contribute  to  the  professional  growth  of  their  colleagues  is  a  direct  responsibility  of  school  leaders.   Global  research  shows  that  school  directors  have  a  large  impact  on  teacher  quality,  both  by  screening   and  selecting  high-­‐talent  teachers  for  their  schools  and  by  engendering  a  school  climate  of  peer   collaboration,  supportive  feedback,  and  collaboration  that  makes  those  teachers  even  better  (Loeb,   Kalogrides  and  Beteille  2012).  High-­‐performing  education  systems  such  as  Singapore  and  Ontario   (Canada)  pay  close  attention  to  how  school  directors  are  selected,  trained,  and  developed,  placing       special  emphasis  on  their  ability  to  gauge  and  develop  the  quality  of  their  teachers  (Schwartz  and  Mehta   2014,  Tucker  2011,Barber  and  Mourshed  2007).   Despite  increasing  awareness  of  the  pivotal  role  of  principals,  empirical  evidence  on  how  to  build  their   skills  and  effectiveness  is  sparse.  Most  LAC  countries  are  just  beginning  to  develop  systems  for  the   selection,  training,  and  coaching  of  school  leaders.     Chile  provides  a  good  example  of  an  incremental  strategy  for  raising  the  quality  of  school  leaders.  Just  as   with  teacher  policy,  the  Ministry  of  Education  started  by  defining  standards.  The  “Framework  for  Good   School  Leadership”  (Marco  para  la  Buena  Dirección),  developed  in  2004,  established  criteria  for  the   training  and  assessment  of  principals  in  the  areas  of  leadership,  curriculum  management,  resource   management,  and  organizational  environment  management  and  established  a  competitive  process  for   principal  selection  (Ministerio  de  Educación,  Chile  2005;  Concha  2007).  A  2011  law  strengthened  the   selection  process  and  increased  principals’  autonomy  and  accountability.  Principals  may  dismiss  up  to  5   percent  of  their  schools’  teachers  each  year  on  performance  grounds  and  must  sign  performance   agreements  with  the  local  governments  that  hire  them.  The  government  also  introduced  the  Program   for  the  Training  of  Excellent  Principals  (Programa  de  Formación  de  Directores  de  Excelencia)  in  2011,   which  has  provided  leadership  training  to  more  than  1,600  of  the  country’s  7,000  principals.  The   program  subsidizes  fees  and  subsistence  costs  for  graduate  programs  (master’s  degrees,  diplomas,   courses)  and  externships  focused  on  school  leadership.  Programs  are  selected  through  a  public  call  for   proposals:  in  2013  applicants  could  choose  among  29  programs  from  15  institutions,  mostly  in  Chile  but   also  in  Canada  and  England.     While  Chile’s  approach  allows  for  diversity  in  the  training  offered  to  principals,  several  OECD  countries   have  chosen  to  develop  in-­‐house  principal  training.  Australia’s  2010  Institute  for  Teaching  and  School   Leadership  develops  standards,  accreditation,  and  training  for  teachers  and  school  leaders  (OECD  2012).   Principals  are  also  trained  in-­‐house  in  Singapore,  where  young  teachers  are  evaluated  for  leadership   potential  early  in  their  careers  and  follow  a  specialized  leadership  track.  This  approach  was  adopted   recently  by  Jamaica,  whose  National  Center  for  Educational  Leadership  is  charged  with  training  and   certifying  aspiring  and  existing  principals.   Once  principals  are  selected  and  trained,  ongoing  support  during  the  early  years  is  important.  A  study  of   New  York  City  schools  found  that  effective  support  for  principals  on  the  job,  particularly  in  the  first  few   years,  has  a  significant  positive  effect  on  school  performance  as  measured  by  student  exam  scores  and   student  absenteeism.  (Clark  et  al.  2009)  Leading  countries  such  as  Singapore  ensure  that  experienced   principals  mentor  new  ones  in  a  systematic  manner.     Motivating  Teachers  to  Perform   Major  progress  in  raising  the  quality  of  teachers  in  Latin  America  will  require  attracting  high-­‐caliber   candidates,  continuously  and  systematically  weeding  out  the  lowest  performers,  and  motivating   individuals  to  keep  refining  their  skills  and  working  their  hardest  over  a  long  career.  These  three   processes  characterize  the  labor  market  for  high  status  professions  in  all  countries.  In  countries  with   high-­‐performing  education  systems,  they  operate  in  teaching  as  well.     Research  confirms  that  individuals  are  attracted  to  the  teaching  profession  and  inspired  to  high   performance  for  a  variety  of  reasons.  Vegas  and  Umansky  (2005)  set  out  a  comprehensive  framework  of   incentives  that  may  be  collapsed  into  three  broad  categories:  (a)  professional  rewards,  including  intrinsic   satisfaction,  recognition  and  prestige,  professional  growth,  intellectual  mastery,  and  pleasant  working   conditions;  (b)  accountability  pressure;  and  (c)  financial  incentives  (figure  O.20).         Figure  O.20  Three  broad  classes  of  incentives  motivate  teachers     Source:  Adapted  from  Vegas  and  Umansky  2005.   While  it  seems  intuitively  obvious  that  all  three  types  of  incentives  are  important,  there  is  a  deep   asymmetry  in  the  research  base.  Very  little  research  exists  on  specific  policies  or  programs  to  raise  the   professional  rewards  for  teachers,  and  none  in  Latin  America.  There  is  more  research  on  reforms  to   strengthen  accountability  pressures  on  teachers—especially  through  school-­‐based  management—but   little  evidence  on  key  questions  such  as  the  impact  of  policies  that  reduce  teachers’  job  stability  or   improve  school  directors’  capacity  to  evaluate  and  manage  teacher  performance.  The  greatest  research   attention  by  far  has  been  focused  on  financial  incentives,  especially  bonus  pay.  But  this  research  bias   should  not  be  taken  to  mean  that  financial  incentives  are  the  most  important.  If  anything,  cross-­‐country   studies  suggest  that  professional  incentives  are  a  very  powerful  element  in  high-­‐performing  education   systems.  In  Finland  and  Canada,  for  example,  the  professional  rewards  for  teachers  are  very  strong  and   accountability  pressures  and  financial  rewards  are  relatively  weak.     It  is  also  likely  that  these  three  types  of  incentives  are  complementary:  they  have  extra  impact  if  well-­‐ aligned  and  undercut  each  other  if  not.  Case  studies  of  the  world’s  highest-­‐performing  school  systems   show  positive  incentives  in  all  three  areas,  although  their  relative  strength  can  vary.     Professional  rewards.  There  is  virtually  no  experimental  evidence  on  the  impact  of  alternative  strategies   for  raising  the  professional  rewards  to  teaching,  but  cross-­‐country  studies  show  that  high-­‐performing   school  systems  offer  their  teachers  abundant  opportunities  for  continued  mastery  and  professional   growth  and  that  outstanding  teachers  receive  substantial  recognition  and  prestige.  Compared  with  most   of  Latin  America,  countries  with  high-­‐performing  education  systems  invest  more  resources  in  teacher   professional  development  –  Singapore’s  100  hours  annually  of  paid  professional  development  for  every       teacher  is  a  leading  example  –  but  most  important  is  the  quality  of  those  investments.  Courses  are   developed  by  university  practitioners  in  close  collaboration  with  Ministries  of  Education,  grounded  in   research  evidence,  and  focus  on  specific  issues  in  effective  delivery  of  the  curriculum,  the  classroom   practice  of  highly  effective  teachers,  and  lessons  from  education  systems  elsewhere  in  the  world.  High-­‐ performing  systems  also  support  teachers’  professional  growth  by  promoting  constant  interaction  and   peer  collaboration  among  teachers.  Finland’s  teachers  spend  only  60  percent  as  much  time  as  the  OECD   average  in  the  classroom  teaching;  the  rest  of  their  time  they  work  jointly  on  new  curriculum  content,   learning  materials  and  ways  of  assessing  students’  progress.  Peru’s  teacher  mentoring  program  and  Rio   de  Janeiro  municipality’s  ginasio  carioca  experimental  are  promising  new  examples  in  the  LAC  region  of   efforts  to  promote  teachers’  professional  mastery  through  peer  collaboration.     High-­‐performing  education  systems  also  give  substantial  recognition  and  prestige  to  excellent  teachers.   They  have  systems  to  evaluate  individual  teachers’  potential  and  performance  and  grant  the  best   teachers  special  status  as  master  teachers  or  leaders  in  specific  curriculum  areas,  such  as  math.  In   contrast,  teachers  in  Latin  America  are  rarely  observed  or  evaluated  closely.  Whether  their  performance   is  outstanding  or  deeply  deficient,  teachers  in  most  systems  advance  equally  through  the  ranks  on  the   basis  of  seniority.     Accountability  pressure.  High  teacher  absence  rates  across  the  LAC  region  and  classroom  observations   showing  that  teachers  are  often  poorly  prepared  to  use  class  time  effectively  are  evidence  that  the   pressures  teachers  feel  to  perform  accountably  are  generally  weak.  Strategies  for  strengthening   accountability  include  actions  to  reduce  or  eliminate  teachers’  job  stability,  increase  managerial   oversight,  and  empower  clients  (parents  and  students)  to  monitor  or  evaluate  teachers.  There  is  little   research  evidence  to  date  on  any  of  these  strategies  except  client  empowerment:  “strong”  forms  of   school-­‐based  management,  in  which  parents  and  community  members  had  a  voice  in  the  hiring  and   firing  of  school  personnel  and  were  given  training  and  encouragement  to  exercise  that  power,  has  been   shown  in  some  contexts  to  reduce  teacher  absence  and  raise  student  learning  results  (Bruns  et  al.  2011).     In  terms  of  job  stability,  legislation  in  Chile,  Peru,  Ecuador,  Colombia,  and  Mexico  has  created  a  path  out   of  the  profession  for  teachers  with  persistently  unsatisfactory  performance.  While  potentially  very   important,  the  number  of  teachers  dismissed  to  date  in  these  countries  has  been  tiny.  This  contrasts   with  the  practice  in  Singapore,  where  all  teachers  are  evaluated  regularly  and  the  lowest  5  percent  of   performers  are  counseled  out  of  the  profession  annually,  and  in  Washington,  DC,  in  which  33  percent  of   the  teaching  force  was  either  dismissed  or  left  voluntarily  in  the  first  four  years  after  its  teacher   evaluation  system  was  introduced.    Radical  upgrading  of  the  teaching  profession  in  Latin  America  will   require  much  more  aggressive  action  to  weed  out  the  lowest-­‐performing  teachers  on  a  continuous   basis.     In  terms  of  managerial  oversight,  the  role  of  school  directors  in  managing  teacher  performance  in  LAC   has  generally  been  weak.  Research  in  the  United  States  has  documented  what  many  LAC  education   leaders  observe:  high-­‐performing  schools  achieve  success  through  skilled  management  of  the  teaching   force.  These  schools’  principals  attract  good  teachers,  weed  out  ineffective  teachers,  and  deploy  and   develop  new  teachers  more  effectively  than  principals  at  other  schools.  Effective  principals  have  the   capacity  to  observe  teachers  in  the  classroom,  give  them  formative  feedback,  and  manage  their  “exit”   from  the  school  if  necessary.  Effective  principals  support  teachers’  development  and  hold  them   accountable  for  performance  (Branch  et  al.  2013,  Boyd  et  al.  2008,  Loeb  et  al.  2012).     Recent  efforts  in  Chile,  Jamaica,  Brazil,  Peru,  and  Ecuador  to  raise  the  standards  for  school  directors  and   train  and  empower  these  to  be  accountable  for  instructional  quality  and  teacher  development  are   important  initiatives.  But  Colombia’s  experience—where  directors  found  it  difficult  to  give  teachers   critical  feedback—suggests  that  relying  only  on  school  directors  for  teacher  performance  evaluation  can       be  problematic.  Instead,  teacher  evaluation  needs  to  be  a  systemwide  function  supported  by  expert   external  observers  and  guided  by  common  standards,  evaluation  processes,  and  rubrics,  both  for   fairness  and  for  systemwide  learning.  The  goals  are  to  generate  actionable,  formative  feedback  to   teachers  across  the  system  and  to  weed  out  the  lowest  performers  on  a  continuous  basis  so  the  average   quality  of  the  teaching  force  keeps  rising  over  time.     Financial  incentives.  Cross-­‐country  research  suggests  that  the  financial  rewards  for  teaching  must  meet   a  threshold  level  of  parity  with  other  professions  to  attract  high  talent.  As  chapter  1  shows,  average   salaries  and  the  pay  trajectory  for  teachers  in  some  LAC  countries  are  currently  below  this  threshold.     Across-­‐the-­‐board  salary  increases—which  are  politically  popular  and  easy  to  implement—have  the   potential  to  shift  the  overall  teacher  supply  curve  outward.  But  these  are  inefficient.  For  the  same  fiscal   expenditure,  school  systems  can  achieve  higher  quality  by  raising  average  salaries  through  a  pay  scale   differentiated  by  performance.  This  avoids  overcompensating  weak  performers,  can  keep  overall   pension  liabilities  lower,  and  creates  stronger  incentives  for  the  most  talented  individuals.     The  two  main  strategies  for  differentiated  financial  rewards  are  career  path  reforms  and  bonus  pay.   Career  path  reforms  typically  make  permanent  promotions  contingent  on  teachers’  skills  and   performance  rather  than  on  seniority,  and  expand  salary  differentials  across  different  grades.  The   number  of  LAC  countries  that  have  implemented  career  path  reforms  is  small  but  growing.  Although  it  is   difficult  to  evaluate  such  reforms  rigorously,  as  they  are  almost  always  implemented  systemwide,    by   analogy  with  other  occupations  it  is  likely  that  career  path  reforms  have  more  powerful  selection  effects   than  bonus  pay  on  who  goes  into  teaching.  Career  path  reforms  signal  a  permanent  and  cumulative   structure  of  rewards  for  high  performance,  have  attractive  pension  implications,  and  are  reaped  by   individual  teachers.     Key  lessons  that  can  be  drawn  from  the  experience  with  career  path  reforms  in  LAC  to  date  are:     • Choosing  valid  measures  of  teacher  quality  and  calibrating  them  appropriately  are  crucial  steps.   Global  research  suggests  that  comprehensive  teacher  evaluations  are  the  soundest  basis  for   promotion  decisions.  An  example  consistent  with  global  best  practice  is  Peru’s  2012  teacher   career  law,  which  establishes  teacher  knowledge  and  skills,  rather  than  seniority,  as  the  basis  for   promotions.  Ministry  proposals  for  implementing  the  law  call  for  comprehensive  evaluations  of   teacher  quality,  including  expert  observations  of  teachers’  practice  in  the  classroom;  and  “360-­‐ degree”  feedback  from  peers,  students,  parents,  and  school  directors  –  all  consistent  with  the   best  global  evidence.     For  school  systems  introducing  competency-­‐based  promotion  and  pay  for  the  first  time,  relying   on  a  well-­‐designed  test  of  subject  matter  mastery  and  pedagogical  knowledge  alone,  as  in   Ecuador,  may  be  a  practical  first  step.  To  be  legitimate,  tests  should  measure  what  teachers   know  in  terms  of  content,  what  they  understand  about  child  development  and  learning  styles,   and  what  they  are  able  to  do  to  tailor  pedagogical  strategies  for  the  delivery  of  content  at   different  grade  levels.  Tests  must  also  be  benchmarked  appropriately;  if  promotions  are  gained   too  easily,  as  in  the  early  years  of  Mexico’s  Carrera  Magisterial,  or  are  too  inaccessible,  incentive   strength  erodes.  Finally,  the  most  recent  career  path  reforms  in  LAC  do  not  base  teacher   promotions  on  student  test  scores.  This  appears  sensible,  given  the  technical  complexity  of   value-­‐added  learning  measures  and  the  risks  of  perverse  incentives.     • Who  evaluates  is  important.  Although  ministries  of  education  should  closely  control  the  design   and  implementation  of  teacher  promotion  policies,  contracting  external  agencies  to  design  and   administer  teacher  evaluations  increases  their  legitimacy.  For  classroom  observations,  using       well-­‐trained  external  experts,  developing  clear  and  consistent  evaluation  standards  and   instruments,  and  providing  teachers  with  detailed  individualized  feedback  are  important.     • The  steepness  of  the  salary  trajectory  affects  incentive  strength,  but  there  is  little  evidence  as   yet  to  guide  reform  design.  Recent  reforms  expand  the  number  of  promotion  levels  and   decompress  the  band  between  top  and  initial  salaries.  But  across  the  new  programs,  these   dimensions  vary:  three  different  promotion  levels  are  proposed  in  some  systems  and  eight   levels  in  others.  Top-­‐level  salaries  are  100  percent  higher  than  starting  level  in  some  systems,   and  almost  300  percent  higher  in  others.  As  most  of  these  reforms  are  quite  new,  there  is  an   important  opportunity  to  research  their  differential  impacts  on  the  recruitment  of  new  teachers   over  time.     • Strategies  for  managing  the  long-­‐term  fiscal  implications  of  career  path  reforms  are  important.   Although  the  permanence  of  promotions  and  base  pay  increases  is  central  to  its  strength  as  an   incentive,  it  runs  the  risk  of  locking  in  high  compensation  for  teachers  who  are  promoted  but   subsequently  fail  to  keep  up  their  skills.  Ecuador’s  reform  guards  against  this  by  requiring  that   teachers  achieve  either  further  promotion  or  recertification  at  the  same  level  every  four  years,   or  face  a  downgrade  of  level  and  salary.  This  is  the  first  case  in  the  region  of  an  “up  or  out”   strategy  built  into  a  career  path  reform.  It  is  an  interesting  design  that  deserves  evaluation.     • Careful  implementation  planning  of  reforms  as  complex  as  these  is  important.  The  credibility  of   several  programs—Mexico’s  original  Carrera  Magisterial,  Peru’s  2008  Carrera  Publica   Magisterial  reform,  and  programs  in  Colombia  and  Sao  Paulo—has  been  undermined  by   problems  that  might  have  been  foreseen  and  managed  differently.     • Incentive  power  hangs  on  belief  that  the  program  will  be  sustained  under  consistent  rules  of  the   game.  Whenever  teachers  perceive  that  the  criteria  for  entry  into  a  new  career  track  are  likely   to  change,  soften,  or  be  disbanded,  the  incentives  to  acquire  new  knowledge  and  apply  it  to   their  work  erode.  Career  path  reforms  that  truly  signal  substantially  higher  long-­‐term  financial   rewards  for  talented  teachers  probably  offer  the  clearest  path  to  the  recruitment  of  higher-­‐ caliber  teacher  candidates  and  more  effective  teaching.  Policymakers  across  the  region  would   gain  from  careful  research  on  the  new  wave  of  career  path  reforms  in  LAC.   Bonus  pay.  Bonus  pay  is  the  other  major  instrument  for  raising  the  financial  rewards  for  teaching.  Bonus   pay  programs  are  proliferating  in  LAC,  especially  in  Brazil.  They  are  politically  and  technically  easier  to   implement  than  career  path  reforms  and  do  not  have  long-­‐term  fiscal  or  pension  implications.  Bonus   programs  typically  offer  a  one-­‐time  reward  for  teachers  (or  schools)  for  specific  results  achieved  during   the  prior  school  year.  There  is  no  evidence  yet  on  the  impact  of  bonus  pay  programs  on  the  critical  long-­‐ term  question  of  teacher  selection:  are  bonus  pay  programs  a  sufficiently  strong  financial  incentive  to   attract  higher-­‐caliber  candidates  into  teaching?  But  the  experience  to  date  provides  some  evidence  of   short-­‐term  impacts  on  teacher  and  school  performance  and  lessons  for  program  design:   • Bonus  pay  programs  can  work  in  developing  country  contexts.  Although  the  number  of  cases   remains  small,  bonus  pay  programs  in  developing  country  settings  have  produced  more   consistently  positive  results  than  in  developed  countries  (especially  the  United  States)  to  date.   The  only  two  rigorously  evaluated  cases  of  bonus  pay  programs  operating  at  scale  (Chile’s  SNED     and  the  school  bonus  in  Pernambuco,  Brazil)  are  both  from  Latin  America,  and  both  have   demonstrated  positive  results  on  student  learning  and  grade  attainment.  Measured  impacts   across  all  developing  country  programs  to  date  are  generally  in  the  range  of  0.10  to  0.3  SD   improvements  in  test  scores,  which  are  significant-­‐sized  effects  for  education  interventions.  A   reasonable  hypothesis  is  that  bonus  pay  incentives—which  focus  schools  on  student  learning       results—can  be  productive  in  systems  where  other  accountability  pressures  and  teacher   professionalism  are  weak.     • Matching  incentive  design  to  context  is  crucial.  Much  of  the  experimental  evidence  to  date  is   from  studies  that  tested  alternative  bonus  designs—group  versus  individual  incentives;  teacher   versus  student  incentives;  “gain”  versus  “loss”  bonus  awards—and  it  is  striking  how  much  the   impact  of  alternative  bonus  designs  can  vary  within  a  single  context.  The  optimal  bonus  size  is   another  design  issue  on  which  there  is  as  yet  little  practical  guidance  from  research;  some  of  the   largest  reported  impacts  in  the  literature  are  from  bonuses  that  represented  a  very  small   increment  of  teachers’  monthly  pay,  but  significantly  larger  bonuses  -­‐-­‐  averaging  1-­‐2  months’   salary  -­‐-­‐  are  becoming  common  in  Brazil.  The  research  base  today  is  far  short  of  providing  a   guide  to  the  most  productive  bonus  pay  designs  for  a  given  context.  But  it  suggests  that  if  a   given  program’s  impact  appears  to  be  weak,  there  probably  exists  a  productive  alternative   design.       • Designing  the  performance  measure(s)  to  be  rewarded  is  a  key  challenge.  Basing  bonus  pay  on   student  test  scores  alone  has  been  problematic  in  several  US  settings  because  of  documented   cheating  and  broader  concerns  that  it  focuses  teachers  too  narrowly  on  test  preparation  and   specific  subjects,  and  makes  them  unwilling  to  teach  at-­‐risk  students.  No  LAC  country  to  date   has  introduced  bonus  pay  based  on  test  scores  alone,  and  this  strategy  appears  wise.  The   composite  indicator  used  in  Brazil,  which  is  a  product  of  test  scores  and  pass  rates,  is  an   interesting  model  for  countries  to  consider.  It  discourages  automatic  promotion  of  children  who   are  not  learning  and  the  reverse  strategy  of  holding  children  back  or  encouraging  dropout  to   boost  test  scores.     • Programs  may  have  heterogeneous  impacts  on  different  types  of  schools.  In  both  of  the  bonus   programs  operating  at  scale,  significant  heterogeneity  has  been  observed.  In  Chile’s  SNED   program,  about  one-­‐third  of  schools  appear  consistently  “out  of  the  money”  in  the  bonuses   granted  every  two  years,  despite  the  serious  efforts  made  to  ensure  that  schools  compete  only   against  similar  schools  (Contreras  and  Rau  2012).  In  the  case  of  Pernambuco,  Brazil,  the  bonus   has  produced  stronger  improvements  in  small  schools,  where  teachers  can  collaborate  and   monitor  each  other  more  easily  than  in  larger  schools.  Improvements  have  also  been  larger  for   academically  weak  and  low-­‐income  students,  suggesting  that  the  bonus  has  stimulated  schools   and  teachers  to  focus  more  effort  on  these  students  (Ferraz  and  Bruns  2014).  Research  evidence   of  this  type  can  provide  useful  guidance  for  program  design.       • Students  are  a  key  partner  in  the  production  of  learning  results.  The  innovative  design  of  the  ALI   (Aligning  Learning  Incentives)  experiment  in  Mexico  generated  powerful  evidence  that  school   systems  can  gain  by  finding  ways  to  make  students  feel  more  invested  in  their  learning  progress   (Behrman  et  al.  2014).  This  is  consistent  with  evidence  that  student  learning  performance  on   international  tests  is  higher  for  countries  with  high-­‐stakes  examinations  for  students  at  the  end   of  secondary  school,  which  create  strong  incentives  for  student  effort  (Woessmann  2012).   • Our  understanding  of  the  mechanisms  through  which  bonus  pay  improves  student  outcomes  is   still  weak.  The  logic  of  incentive  pay  is  to  stimulate  teacher  behaviors  that  help  raise  student   learning:  either  increased  teacher  effort  or  more  effective  effort.  However,  relatively  few   evaluations  have  documented  changes  in  teachers’  classroom  practice  that  plausibly  explain   observed  increases  in  student  learning.  Research  on  teachers’  classroom  practice  is  becoming   more  feasible,  with  the  declining  costs  of  installing  video  cameras  in  samples  of  classrooms  and   increasing  use  of  standardized  methods  for  coding  and  analyzing  teacher-­‐student  interaction.       Systematic  inclusion  of  such  analysis  in  impact  evaluations  of  pay  for  performance  programs  will   not  only  illuminate  how  such  programs  work  but  also  generate  evidence  and  examples  of   effective  teaching  that  can  more  broadly  benefit  these  school  systems.     Ultimately,  cross-­‐country  studies  suggest  that  no  education  system  achieves  high  teacher  quality   without  aligning  all  three  types  of  incentives:  professional  rewards,  accountability  pressures,  and   financial  rewards.  But  these  studies  also  suggest  that  the  particular  combinations  that  are  most  efficient   are  highly  context-­‐specific.  Finland,  Singapore,  and  Ontario,  Canada,  for  example,  have  all  built  strong   professional  rewards  for  teaching,  but  accountability  pressures  are  much  stronger  in  Singapore  than  in   Finland  or  Canada.  And  none  follows  a  textbook  approach  on  financial  incentives:  Finland  has  achieved  a   sharp  upgrading  of  teacher  quality  over  the  past  20  years  with  little  increase  in  teachers’  relative   salaries.  Singapore  keeps  entering  salaries  for  teachers  on  par  with  other  professions  and  offers  bonuses   for  high  performance,  but  has  an  overall  career  ladder  that  is  much  flatter  than  in  other  professions.   Ontario  pays  competitive  salaries,  but  the  core  of  its  strategy  is  team-­‐based  professional  development   at  the  school  level  supported  by  outside  experts  but  not  otherwise  incentivized.  These  examples  suggest   that  there  are  multiple  roads  to  the  goal:  a  balanced  set  of  incentives  sufficient  to  attract  talented   teacher  candidates,  establish  accountability  for  results,  and  motivate  continued  professional  growth  and   pursuit  of  excellence.     Managing  the  Politics  of  Teacher  Reform     Teachers  are  not  only  key  actors  in  the  production  of  education  results  but  also  the  most  powerful   stakeholders  in  the  process  of  education  reform.  No  other  education  actor  is  as  highly  organized,  visible,   and  politically  influential  (Grindle  2004).  Because  of  their  unique  autonomy  behind  the  closed  door  of   the  classroom,  teachers  also  have  profound  power  over  the  extent  to  which  new  policies  can  be   implemented  successfully.  By  global  standards,  teachers’  unions  in  Latin  America  and  the  Caribbean   have  been  considered  especially  powerful.  They  have  a  history  of  effective  use  of  direct  electoral   influence  and  disruptive  actions  in  the  streets  to  block  reforms  perceived  as  a  threat  to  their  interests.     Like  all  organized  workers,  teachers’  unions  exist  to  defend  the  rights  they  legitimately  earn  through   negotiations  and  to  oppose  policy  changes  that  threaten  those  rights.  Teachers  and  their   representatives  are  entirely  justified  in  pursuing  these  goals,  and  teachers’  unions  throughout  history   have  been  a  progressive  force  in  achieving  equal  pay  and  fair  treatment  for  women  and  minority   members.  But  it  is  also  true  that  the  goals  of  teachers’  organizations  are  not  congruent  with  the  goals  of   education  policy  makers  or  the  interests  of  education  beneficiaries—including  students,  parents,  and   employers  who  need  skilled  workers.       Viewed  through  the  lens  of  teachers’  legitimate  interests,  a  number  of  education  policies  that   goverments  adopt  in  pursuit  of  education  quality  pose  threats:    to  teachers’  benefits  (elimination  of  job   stability  and  reduction  or  loss  of  other  benefits);  to  teachers’  working  conditions  (curriculum  reforms,   student  testing,  teacher  evaluation  systems);  or  to  union  structure  and  power  (decentralization,  school   choice,  higher  standards  for  teachers  at  entry,  alternative  certification,  and  pay  linked  to  individual  skills   or  performance).    Relatively  few  education  policies—higher  spending  on  education,  bonus  pay  at  the   school  level,  and  lower  pupil-­‐teacher  ratios—are  positively  aligned  with  unions’  interests.  Unions’  ability   to  challenge  policies  depends  on  their  structure  (i.e.,  share  of  teachers  unionized),  their  capacity  for   collective  action,  and  the  effectiveness  of  their  political  strategies.  The  latter  include  strikes  and   protests,  government  capture,  legal  strategies,  and  union-­‐sponsored  research  and  policy  analysis  to   influence  education  debate.  All  of  these  strategies  have  been  deployed  effectively  by  unions  in  Latin   America  and  the  Caribbean  in  national  debates  over  education  reform  during  the  past  several  decades.         But  recent  reform  experiences  in  Mexico,  Peru,  and  Ecuador  suggest  that  the  balance  of  power  between   governments  and  teachers’  unions  in  the  region  may  be  shifting.  In  an  age  where  mass  media  afford   political  leaders  a  direct  channel  of  communication  with  even  the  most  remote  and  rural  of  their   citizens,  one  of  the  long-­‐time  sources  of  union  power—the  ability  to  mobilize  their  members  for  large-­‐ scale  grassroots  political  campaigning—may  be  of  diminishing  utility.  In  a  region  where  democracy  has   taken  hold  in  most  countries,  mass  media  have  become  increasingly  vociferous  in  exposing  government   failure  and  political  corruption.  This  feeds  public  demand  for  more  accountable  and  effective   government  and  resonates  particularly  strongly  in  education,  which  touches  every  family’s  hopes  and   aspirations  for  its  children.  Increasingly,  political  leaders  in  LAC  appear  to  be  calculating  that  popular   support  for  education  reform  is  a  stronger  bet  for  their  political  future  than  the  traditional  quid  pro  quo   of  electoral  support  from  teachers’  unions  in  exchange  for  education  policies  that  do  not  threaten  their   interests.   While  there  is  substantial  heterogeneity  across  the  region  in  union  power,  government  reform  priorities,   and  the  dynamics  of  the  reform  process,  the  most  recent  reform  experiences  support  several  cautious   observations:   • Political  leaders  can  build  effective  proreform  alliances  of  business  leaders  and  civil  society   through  communications  campaigns  that  paint  a  compelling  picture  of  the  current  failures  of  the   education  system  and  the  importance  of  better  education  for  economic  competitiveness.   Successfully  uniting  two  sides  of  the  stakeholder  triangle  (civil  society  and  government)  in   dialogue  with  the  third  (organized  teachers)  can  create  political  space  for  the  adoption  of   reforms,  including  three  that  challenge  union  interests  (individual  teacher  performance   evaluation,  pay  differentiated  by  performance,  and  loss  of  job  stability).     • Reform  momentum  is  greatest  if  launched  at  the  start  of  an  administration.  In  most  cases,  the   process  is  contentious  and  unions  have  a  strong  interest  in  dragging  it  out.  If  leaders  move   quickly,  they  capitalize  on  their  point  of  maximum  political  leverage  and  establish  education  as  a   top  priority.  As  they  begin  to  govern,  administrations  are  inevitably  forced  to  spend  time  on  a   wide  range  of  other  issues  and  suffer  some  political  reversals;  this  diffuses  messages  and  erodes   leverage.       • Hard  data  on  education  system  results  are  a  crucial  political  tool.  Especially  powerful  are  data  on   student  learning  outcomes,  results  that  are  internationally  benchmarked  (such  as  PISA,  TIMMS,   SERCE,  and  LLERCE  and  data  on  teachers’  performance  on  competency  tests.  Political  leaders’   use  of  these  to  build  the  case  for  reform  has  been  a  factor  in  all  successful  strategies  to  date.  Of   all  international  tests,  the  OECD’s  PISA  seems  to  resonate  most  strongly  with  the  business   community  and  civil  society  groups.  This  is  likely  because  the  comparator  countries  are  those   that  LAC  countries  aspire  to  join,  and  because  it  is  easy  to  interpret  the  results,  for  15-­‐year-­‐old   youths,  as  a  barometer  of  labor  force  quality  and  economic  competitiveness.       • Reform  strategies  based  on  confrontation  with  unions  may  succeed  in  securing  the  legislative   adoption  of  major  reforms,  but  not  necessarily  their  implementation.  In  many  countries,  the   political  space  for  negotiating  major  reforms  with  teachers’  unions  does  not  exist.  In  three   recent  cases  (Mexico,  Peru,  and  Ecuador)  confrontation  politics  has  produced  the   legislative/constitutional  adoption  of  teacher  policy  reforms  that  global  evidence  suggests  are   needed  for  education  quality:  student  testing,  teacher  performance  evaluation,  teacher  hiring   and  promotions  linked  to  skills  and  performance  rather  than  seniority,  and  dismissal  of  teachers   with  consistently  poor  performance.  There  may  be  no  political  alternative  to  confrontation   strategies  in  many  contexts;  in  Mexico’s  case,  a  high-­‐profile  government  effort  to  design       reforms  in  collaboration  with  the  union  foundered  when  the  union  could  not  deliver  members’   adherence  with  the  agreements..  But  confrontation  strategies  imply  a  major  trade-­‐off:  they   make  it  impossible  to  gain  input  from  teachers  that  could  genuinely  improve  a  reform’s  design   and  smooth  its  implementation.     • Sequencing  reforms  can  ease  adoption  and  improve  implementation.  The  region’s  experience   suggests  a  political  logic  to  a  certain  sequence  of  education  reforms.  The  first  step  is  student   testing,  with  transparent  dissemination  of  results,  both  nationally  and  to  individual  schools;  this   is  the  anchor  that  makes  it  possible  to  introduce  other  performance-­‐based  reforms.  A  second   step  is  the  adoption  of  school-­‐based  bonus  pay,  which  establishes  the  concept  of  pay  for   performance  and  focuses  schools  on  student  learning  progress,  but  has  typically  faced  less   union  resistance  than  individual  bonus  pay.  A  third  step  is  individual  teacher  evaluation  on  a   voluntary  basis,  with  the  carrot  of  attractive  financial  rewards  for  teachers  who  take  the  risk  of   being  evaluated  and  perform  well.  Unions  have  typically  opposed  this,  but  making  programs   voluntary  can  avoid  confrontation.  This  sequence  of  reforms  was  implemented  in  Chile  between   1995  and  2004,  more  recently  by  Sao  Paulo  state,,  and  (proposed)  in  Rio  de  Janeiro  state.   All  of  the  available  evidence  suggests  that  the  quality  of  teachers  in  Latin  America  and  the  Caribbean  is   the  binding  constraint  on  the  region’s  progress  toward  world-­‐class  education  systems.  Low  standards  for   entry  into  teaching;  low-­‐quality  candidates;  salaries,  promotions,  and  job  tenure  delinked  from   performance;  and  weak  school  leadership  have  produced  low  professionalism  in  the  classroom  and  poor   education  results.  Moving  to  a  new  equilibrium  will  be  difficult,  and  will  require  recruiting,  grooming,   and  motivating  a  new  breed  of  teacher.     Raising  the  stakes  further  are  the  sweeping  transformations  occurring  in  global  education.  The   traditional  goals  of  national  education  systems  and  the  traditional  paradigm  of  teacher-­‐student   interaction  made  teachers  the  linchpin  in  the  transmission  of  discrete  bodies  of  knowledge  to  students   in  the  classroom.  The  new  paradigm  is  that  teachers  are  not  the  only  or  even  the  major  source  of   information  and  knowledge  available  to  students.  A  core  role  of  teachers  today  is  to  equip  students  to   seek,  analyze,  and  effectively  use  vast  amounts  of  information  that  are  readily  available  elsewhere.   Teachers  must  also  develop  students’  competencies  in  the  broad  range  of  areas  valued  in  an  integrated   global  economy:  critical  thinking;  problem-­‐solving;  working  collaboratively  in  diverse  environments;   adapting  to  change;  and  the  capacity  to  master  new  knowledge,  skills,  and  changing  employment   demands  across  their  lifetimes.  No  teacher  preparation  programs  in  LAC—or  indeed  in  most  OECD   countries—are  fully  prepared  to  produce  this  profile  of  teacher  today,  let  alone  the  profiles  that  may  be   needed  over  the  next  decade.  But  virtually  all  OECD  countries  are  responding  to  these  challenges  by   raising  their  expectations,  and  standards,  for  teachers.   Countries  across  Latin  America  and  the  Caribbean  are  also  responding.  Virtually  all  aspects  of  teacher   policy  are  under  review  and  reform  in  different  LAC  countries,  and  in  some  areas  the  region  is  in  the   vanguard  of  global  policy  experience.  By  drawing  together  in  one  volume  the  key  teacher  policy  reforms   being  undertaken  in  the  region  today  and  the  best  available  evidence  on  their  impact,  this  book  hopes   to  stimulate  and  support  the  faster  progress  that  is  needed.   Note   1.  Because  Peru  did  not  participate  in  the  2003  and  2006  rounds  of  PISA,  researchers  excluded  it  from   the  analysis  of  countries  registering  the  most  significant  sustained  progress  between  1990  and  2006.             References   Alfonso,  M.,  A.  Santiago,  and  M.  Bassi.  2010.  Estimating  the  Impact  of  Placing  Top  University   Graduates  in  Vulnerable  Schools  in  Chile.  Technical  Note  IDB-­‐TN-­‐230.  Inter-­‐American   Development  Bank,  Washington,  DC.   Barber,  M.,  and  M.  Mourshed.  2007.  How  the  World’s  Best-­‐Performing  School  Systems  Come  Out  on   Top.  London:  McKinsey.   http://mckinseyonsociety.com/downloads/reports/Education/Worlds_School_Systems_Final .pdf.   Behrman,  J.,  S.  Parker,  P.  Todd,  and  K.  Wolpin.  2014.  “Aligning  Learning  Incentives  of  Students  and   Teachers:  Results  from  a  Social  Experiment  in  Mexican  High  Schools.”  Journal  of  Political   Economy,  forthcoming.     Boyd,  D.,  P.  Grossman,  H.  Lankford,  S.  Loeb,  and  J.  Wyckoff.  2008.  “Who  Leaves?  Teacher  Attrition   and  Student  Achievement.”  Working  Paper  14022,  National  Bureau  of  Economic  Research   Boyd,  D.,  P.  Grossman,  S.  Lankford,  S.  Loeb,  and  J.  Wyckoff.  2009.  “Teacher  Preparation  and  Student   Achievement.”  Educational  Evaluation  and  Policy  Analysis  31  (4):  416–40.   Branch,  G.,  E.  Hanushek,  and  S.  Rivkin.  2013.  “School  Leaders  Matter.”  Education  Next  13  (2):62–69.     Bruns,  B.,  D.  Filmer,  and  H.  A.  Patrinos.  2011.  Making  Schools  Work:  New  Evidence  on  Accountability   Reforms.  Washington,  DC:  World  Bank.   Carnoy,  M.  (2007).  Cuba’s  Academic  Advantage:  Why  Students  in  Cuba  Do  Better  in  School.  Stanford   University  Press.   Chetty,  R.,  J.  N.  Friedman,  and  J.  E.  Rockoff.    2014.  “Measuring  the  Impacts  of  Teachers  II:  Teacher   Value-­‐Added  and  Student  Outcomes  in  Adulthood.”  American  Economic  Review  (forthcoming).       Chingos,  M.,  and  P.  E.  Peterson.  2010.  “Do  Schools  Districts  Get  What  They  Pay  For?  Predicting   Teacher  Effectiveness  by  College  Selectivity,  Experience,  Etc.”  Harvard  University  Program  on   Education  Policy  and  Governance  Working  Paper  10-­‐08,  Harvard  University,  Cambridge,  MA.   Clark,  D.,  P.  Martorell,  and  J.  Rockoff.  2009.  “School  Principals  and  School  Performance.”  CALDER   Working  Paper  38,  National  Center  for  Analysis  of  Longitudinal  Data  in  Education  Research,   Urban  Institute,  Washington,  DC.   Concha  Albornoz,  C.  2007.  “Claves  para  la  formación  de  directivos  de  instituciones  escolares.”   Revista  Electrónica  Iberoamericana  sobre  Calidad,  Eficacia  y  Cambio  en  Educación  5,  (5)  133-­‐ 138.     Contreras,  D.,  and  T.  Rau.  2012.  “Tournament  Incentives  for  Teachers:  Evidence  from  a  Scaled-­‐up   Intervention  in  Chile.”  Economic  Development  and  Cultural  Change  91  (1):  219–46.   Corcoran,  S.  P.,  W.  N.  Evans,  and  R.  M.  Schwab.  2004.  “Women,  the  Labor  Market,  and  the  Declining   Relative  Quality  of  Teachers.”  Journal  of  Policy  Analysis  and  Management  23  (3):  449–470.   (Costa+Carnoy)   Dee,  T.,  and  J.  Wyckoff.  2013.  “Incentives,  Selection  and  Teacher  Performance:  Evidence  from   IMPACT.”  Working  Paper  19529,  National  Bureau  of  Economic  Research.   Eide,  E.  G.,  D.  Goldhaber,  and  D.  Brewer.  2004.  “The  Teacher  Labour  Market  and  Teacher  Quality.”   Oxford  Review  of  Economic  Policy  20  (2):  230–44.   Ferraz,  C.,  and  B.  Bruns.  (forthcoming,  2014).  “Paying  Teachers  to  Perform:  The  Impact  of  Bonus   Pay  in  Pernambuco,  Brazil.”  Manuscript.  ,  World  Bank,  Washington,  DC.   Franco,  M.  2012.  “Pre-­‐Service  Training  in  Latin  America  and  the  Caribbean:  A  Background  Study  for   the  World  Bank  LAC  Study  on  Teachers.”  Manuscript,  Washington,  DC.     Fredriksson,  P.,  and  B.  Ockert.  2007.  “The  Supply  of  Skills  to  the  Teacher  Profession.”  Manuscript.   Uppsala  University,  Uppsala,  Sweden.         Garland,  S.  2008.  “Reform  School.”  Daily  Beast,  December  17.   http://www.thedailybeast.com/newsweek/2008/12/17/reform-­‐school.html.   Grindle,  M.  S.  2004.  Despite  the  Odds:  The  Contentious  Politics  of  Education  Reform.  Princeton,  NJ:   Princeton  University  Press.   Hanushek,  E.A.  2011.  “The  economic  value  of  higher  teacher  quality.”    Economics  of  Education   Review,  30:466-­‐479.   Hanushek,  E.  A.,  and  L.  Woessmann.  2007.  “The  Role  of  Education  Quality  for  Economic  Growth.”   Policy  Research  Working  Paper  4122,  World  Bank,  Washington,  DC.   ———.  2012.  “Schooling,  Educational  Achievement,  and  the  Latin  American  Growth  Puzzle.”   Journal  of  Development  Economics  99  (2):  497–512.   Hanushek,  E.,  and  S.  Rivkin.  2010a.  “Using  Value-­‐Added  Measures  of  Teacher  Quality.”  Policy  Brief  9,   National  Center  for  Analysis  of  Longitudinal  Data  in  Education  Research,  Washington,  DC.   Hanushek,  E.  A.,  and  S.  G.  Rivkin.  2010b.  “Generalizations  about  Using  Value-­‐Added  Measures  of   Teacher  Quality.”  American  Economic  Review  100  (2):  267–71.   Hanushek,  E.,  Peterson,  P.,  &  Woessmann,  L.  (2012).  Achievement  Growth:  International  and  U.S   State  Trends  in  Student  Performance.  Harvard’s  Program  on  Education  Policy  and  Governance   &  Education  Next.   Hernani-­‐Limarino,  W.  2005.  “Are  Teachers  Well  Paid  in  Latin  America  and  the  Caribbean?  Relative   Wage  and  Structure  of  Returns  of  Teachers.”  In  Incentives  to  Improve  Teaching:  Lessons  from   Latin  America,  edited  by  E.  Vegas.  Washington,  DC:  World  Bank.   Hoxby,  C.  M.,  and  A.  Leigh.    2004.  “Pulled  Away  or  Pushed  Out?  Explaining  the  Decline  of  Teacher   Aptitude  in  the  United  States.”  American  Economic  Review  94  (2):  236–40.       Hoxby,  C.  M.  (2004).  Pulled  Away  or  Pushed  Out?  Explaining  the  Decline  of  Teacher  Aptitude  in  the   United  States.  American  Economic  Review  ,  94(2):  236–240.   Kane,  T.  J.,  and  D.  O.  Staiger.  2012.  Gathering  Feedback  for  Teaching:  Combining  High-­‐Quality   Observations  with  Student  Surveys  and  Achievement  Gains.  Seattle:  Measures  of  Effective   Teaching  project,  Bill  &  Melinda  Gates  Foundation.   Loeb,  S.,  D.  Kalogrides,  and  T.  Béteille.  2012.  “Effective  Schools:  Teacher  Hiring,  Assignment,   Development,  and  Retention.”  Education  Finance  and  Policy  7  (3):  269–304.   McEwan,  P.  J.,  E.  Murphy-­‐Graham,  D.  Torres  Irribarra,  C.  Aguilar,  and  R.  Rápalo,  R.  Forthcoming.   “Improving  Middle  School  Quality  in  Poor  Countries:  Evidence  from  the  Honduran  Sistema  de   Aprendizaje  Tutorial.”  Educational  Evaluation  and  Policy  Analysis.   Ministerio  de  Educación,  Chile.  2005.  Marco  para  la  Buena  Dirección.  Santiago:  MINEDUC.   http://www.mineduc.cl/usuarios/convivencia_escolar/doc/201103070155490.MINEDUC.M arco_para_la_Buena_Direccion.pdf   Ministerio  de  Educación,  Chile.  2008.  “Marco  para  la  Buena  Enseñanza.”  Docente  más.  Santiago.   (accessed  July  24,  2012),  http://www.docentemas.cl/docs/MBE2008.pdf.   Ministerio  de  Educacion  del  Ecuador.  2012.  Sistema  Integral  de  Desarrollo  Profesional  Educativo.   Ministerio  de  Educacion  del  Ecuador,  Quito.  (accessed  September  7,  2012),   http://sime.educacion.gob.ec/Modulo/SIPROFE/index.php?mp=9_0.   Mizala,  A.,  and  H.  Nopo.  2011.  “Teachers’  Salaries  in  Latin  America:  How  Much  Are  They  (Under  or   Over)  Paid?”  IZA  Discussion  Paper  5947,  Institute  for  the  Study  of  Labor,  Bonn.   Mourshed,  M.,  C.  Chijioke,  and  M.  Barber.  2010.  How  the  World's  Most  Improved  School  Systems  Keep   Getting  Better.  London:  OECD  McKinsey.  Available  at       http://mckinseyonsociety.com/downloads/reports/Education/How-­‐the-­‐Worlds-­‐Most-­‐ Improved-­‐School-­‐Systems-­‐Keep-­‐Getting-­‐Better_Download-­‐version_Final.pdf   OECD.  2012.  Preparing  Teachers  and  Developing  School  Leaders  for  the  21st  Century:  Lessons  from   Around  the  World.  Paris:  OECD.   OECD  (Organisation  for  Economic  Co-­‐operation  and  Development).  2013.  Vol.  I  of  PISA  2012   Results:  What  Students  Know  and  Can  Do—Student  Performance  in  Mathematics,  Reading  and   Science.  Paris:  OECD  Publishing.  DOI  :10.1787/19963777.   Ome,  A.  2012.  “The  Effects  of  Meritocracy  for  Teachers  in  Colombia”  (unpublished  manuscript).   Bogota,  Colombia:  Centro  de  Investigación  Económica  y  Social:  Fedesarrollo.     Picardo  J.  O.  2012.  “La  formación  de  docentes  en  América  Latina  y  El  Caribe:  caso  El  Salvador.”   Manuscript.  San  Salvador.       Rockoff,  J.  E.  2004.  “The  Impact  of  Individual  Teachers  on  Student  Achievement:  Evidence  from   Panel  Data.”  American  Economic  Review  94  (2):  247–52.   Rodríguez,  A.,  Dahlman,  C.  J.,  &  Salmi,  J.  (2008).  Knowledge  and  innovation  for  competitiveness  in   Brazil.  World  Bank.   Salmi,  J.  (2009).  The  challenge  of  establishing  world  class  universities.  World  Bank  Publications.   Schwartz,  J.,  and  J.  Mehta.  2014.  “Ontario:  Harnessing  the  Skills  of  Tomorrow.”  In  Lessons  from  PISA   for  Korea,  Strong  Performers  and  Successful  Reformers  in  Education,  edited  by  OECD  Paris:   OECD  Publishing.  Available  at:  http://dx.doi.org/10.1787/9789264190672-­‐en.   Stallings,  J.,  &  Knight,  S.  2003.  “Using  the  Stallings  Observation  System  to  Investigate  Time  on  Task   in  Four  Countries.”  Unpublished  paper  for  the  International  Time  on  Task  (ITOT)  Project,   World  Bank,  Washington,  DC.   Tatto,  M.  T.,  Ingvarson,  L.,  Schwille,  J.,  Peck,  R.,  Senk,  S.  L.,  &  Rowley,  G.  (2008).  Teacher  Education   and  Development  Study  in  Mathematics  (TEDS-­‐M):  Policy,  Practice,  and  Readiness  to  Teach       Primary  and  Secondary  Mathematics.  Conceptual  Framework.  International  Association  for  the   Evaluation  of  Educational  Achievement.  Herengracht  487,  Amsterdam,  1017  BT,  The   Netherlands.   Thames,  M.  H.,  &  Ball,  D.  L.  (2010).  What  Math  Knowledge  Does  Teaching  Require?.  Teaching   Children  Mathematics,  17(4),  220-­‐229.   Tucker,  M.  ed.  2011.  Surpassing  Shanghai:  An  Agenda  for  American  Education  Built  on  the  World’s   Leading  Systems.  Cambridge,  MA:  Harvard  Education  Press.   UNESCO.  2012.  Antecedentes  y  Criterios  para  la  Elaboración  de  Políticas  Docentes  en  América  Latina   y  el  Caribe.  Santiago:  UNESCO.     UNESCO  Institute  for  Statistics.  2010,  2011,  2012.  Online  database  of  UIS  Statistics.  Montreal,   UNESCO  Institute  for  Statistics.  (accessed  January  15,  2013),  http://data.uis.unesco.org/.   Vaillant,  D.,  and  C.  Rossel.  2006.  Maestros  de  escuelas  básicas  en  América  Latina:  Hacia  una   radiografía  de  la  profesión.  Santiago:  Programa  de  Promoción  de  la  Reforma  Educativa  en   América  Latina  y  el  Caribe  (PREAL).   http://www.oei.es/docentes/publicaciones/maestros_escuela_basicas_en_america_latina_pre al.pdf.     Vegas,  E.  (2007).  Teacher  Labor  Markets  in  Developing  Countries.  The  Future  of  Children  Journal,   Vol.  17,  No.1  Spring  ,  219-­‐232.   Vegas,  E.,  and  I.  Umansky.  2005.  “Improving  Teaching  and  Learning  through  Effective  Incentives.”  In   Incentives  to  Improve  Teaching:  Lessons  from  Latin  America,  edited  by  E.  Vegas,  21–62.   Washington,  DC:  World  Bank.       Woessmann,  L.  (2012).  Peering  over  the  Hedge:  How  Do  the  Neighbours  Do  It?    In  CESifo   Forum  (Vol.  13,  No.  3,  pp.  16-­‐20).  Ifo  Institute  for  Economic  Research  at  the  University  of   Munich.