Econ Develop. Countries Exam 1

The Economic Lives of the Poor

• Author Abhijit Banerjee
• India
• 13 countries
• Living standards Measurement surveys
• Large Findings:
• Poor can still make purchasing decisions
• Problems?
• trouble identifying poor (price differences)
• purchasing power, exchange rates do not prove poverty line, reporting periods vary

Many Children Left Behind? Textbooks and Test scores in kenya

• Authors: Glewwe, Paul, Michael Kremer and Sylvie Moulin
• School assistance program (SAP):
• year 1 : 25 schools get textbooks - 75 do not (control)
• year 2: 25 schools get textbooks 25 get grant and 25 control
• researchers administer standardized tests to compare across groups
• Results:
• Control for everything
• Textbooks do not have an effect
• Randomization is important - seems like textbooks would helps but there is an omitted variable
• OVB possibilities - Language Barrier, bad textbooks
• only program impact group - best students did better with textbooks

Causal Effect of Health on Labor Market Outcomes (Experimental Evidence

• H : strong positive association between health and economic prosperity (Healthier workers less susceptible to disease and are more productive)
• Causal effect of Iron deficiency on economic success (Adults):
• Experimental groups receives iron supplement, control receives placebo
• labor market outcomes compared for next 6 months
• Results: support hypothesis, productivity increases in experiment group, increased work capacity (bike)
• Benefits much more muted in woman than in men
• Potential issues:
• potential selection in attrition (non-random attrition
• Sharing doses with family
• Difficulty detecting/measuring effects in some fields
• Overall: evidence effects (somewhat small), great example of carefully run large sample experiment

Free Distribution or Cost sharing? evidence from a randomized Malaria Prevention Exeriment

• Author; Jessica Cohen and Pascaline Dupas
• Health product demand is price sensitive (price: up demand: down)
• Sunk cost effect: perceived value of product is from price
• Pregnant women targeted (do they want a bed net offered at different prices to randomly assess how it is valued
• If you came and showed money for the net you were offered a lower price in an envelope (either same or reduced price)
• Groups:
• 1. Women that would pay higher price and did
• 2. Women that would pay a high price but paid low
• 3 Women that would pay a low price and did
• Group 1 vs Group 2 measures sunk cost
• Group 1 vs Group 3 combined effect
• Both effects found to be zero selection assumed to be zero
• 3 important results
• Demand Curve is steep
• Usage is no lower when price is lower
• no evidence of selection effet or sunk cost effect charging a small positive price is bad, people do not value the good

Schooling and labor Market Consequences of School Construction in Indonesia

• Author: Esther Duflo
• Constructing Primary Schools: INPRES program built schools
• Govt recruits teachers and pays their salaries (teacher/student ration increases)
• Districts measured by intensity (more schools = higher intensity)
• high intensity districts likely started out doing worse (more students less schools/teachers)
• 2 sources of variation:
• Difference between districts where many schools were built compared to fewer
• difference across ages
• natural "difference in difference"
• it is assumed difference between H and L districts remain constant in absence of program
• Results;
• Younger cohorts catching up to older cohorts - gap between age getting smaller
• education levels are increasing over time faster in H districts than L districts (Difference between H and L smaller)
• Biggest effect in poorest region
• sizable return to education in developing country

Worms: Identifying impacts on Education and Helth in the presence of treatment externalities

• Authors: Edward Miguel, Michael Kremer
• parasitic worms infect 1/4 of worlds pop.
• can be effectively treated in 1 dose (constant reinfection from defecating outside, regular treatment probably necessary) (treatment means you cant infect others benefits pop.)
• goods with positive externalities underprovided (people only pay for private benefit, small fraction of social benefit
• Externalities makes effect of program look smaller
• makes estimating impact hard; paper uses tricks to estimate externalities
• gave deworming pill and prevention education to highly infected regions
• collected test scores and attendance to treatment and worm load
• many in treatment group did not recieve treatment (absent due to worm symptoms and dropping out)
• Impact of program: reduced worm load, not behavior
• attrition with the students who did not show up
• percent of students treated did not change much from year to year
• parent waiver caused problems in second year
• spillover occured: if you did not get treatment you were less likely to have worms than those who did
• conclusions
• big impact on attendence, big spillover
• Returns to education because more health, in addition to just heal benefits
• no short run gains in test scores, long run finds positive labor market effects and cognitive gains

Peer Effects, teacher Incentives and the impact of tracking: Evidence from a randomized evaluation in Kenya

• Authors: Esther Duflo, Pascaline Dupas and Michael Kremer
• Tracking - putting high performing students together and low preforming students together (AP, Honors Accelerated...)
• tracking bad? - Worse teachers w worse students?
• beneficial to learn from high performers
• Tracking good? - Teachers more specific in teaching (pace), teachers have small incentives because schools only care about best performers due to entrance exams, otherwise no incentive
• Experimental design: Additional teachers added to 1st grade, half split classes randomly and half used tracking
• some students had civil service teacher (hard to fire, no incentive) or a contract teacher (year to year contract, incentive)
• Groups:
• Tracking ; Contract vs Civil Service
• non-tracking ; Contract vs Civil Service
• control
• Effects of tracking vs no tracking : compare group 1 and 2
• do incentives matter for teachers? (contract better then civil service?)Does tracking effect teacher effort?
• Results: effects found that kids develop better in reading/math in tracked classrooms, giving them a head start when returned to pooled classrooms
• Huge effects of trackig in high and low classrooms
• results persit when tracking ends, kids with randomly better peers do better
• kids just below cutoof dont do any worse
• civil service teachers show up and work more in tacked classroom (no effect for contract teachers)