Miguel and Kremer (2004) Flashcards
primary school deworming project (PSDP) in Kenya
major public health problem in this area
- 93% affected with one worm
- 37% had at least one moderate-heavy infection
treatment group got 2-3 years of additional deworming before control got treated
- choice of school determined by randomisation and not characteristics
RCT design
- school divided by geographic zone (stacked alphabetically)
- listed by ascending school population
- divided into three groups
arbitrary, no way to give preference to one school over another
- also stratifies the school (in each geographic zone, bounds of 1, 2, or 3 in each area)
three types of analysis in the paper
direct treatment effects
- difference between treatment and comparison schools
within-school externality impacts
cross-school externality impacts
health and nutrition impacts
measured in 1999 when group 1 had a year of deworming while group 2 was the control
twice as many in group 2 had moderate/heavy infections
but still not 0% of infections in group 1
- few people refused treatment, but some people were absent when the drug was administered
- absenteeism about 25%
- also potential reinfection
within-school externalities
benefit to being untreated in a treatment school
- better disease environment even if they don’t take treatment
about a 10 percentage point effect (smaller than the 25 percentage point effect of direct treatment)
used to be missing in existing literature
still not ruling out OVB
- always a factor in within-school externalities
cross-school externalities
large reductions in moderate-heavy infections within 3km (2miles) of treatment schools
- again, disease environment might be better
reduction of about 10 percentage points
- similar in magnitude to within-school externalities
educational impacts: school participation
school participation much higher in the treatment group getting deworming
- not necessarily because they come to get the drug
- lower absenteeism in higher grades
reduction in school absenteeism by about 7 percentage points
educational impacts: academic tests
average test gain from deworming is 0
on average, as class sizes get big, test scores fall
- quality of classroom learnign is low
kids that go to school more might be weaker students and find it hard to learn
performance is a result of a whole lifetime of attendance, etc. so time lags until we detect effects
deworming cost-benefit calculations
deworming is very very cheap
healthy behaviours did not appear to have changed in the treatment schools
economic returns to deworming (10 year)
decade after treatment, total school enrolment increased on average with larger gains for females
self-reported health, labour hours, wages all increased
- also changing work occupations (less casual labour and domestic service, more manufacturing)
deworming therefore looks like an attractive investment with large economic returns
longer-term economic returns to deworming (15-20 year)
total earnings, household consumption, household earnings all significantly higher
- total earnings 7 percentage points higher
also higher migration to urban areas
- by about 5 percentage points
strong intergenerational effects
- children much more likely to survive
treatment rates in older girls
WHO had a recommendation that girls of childbearing age should be excluded (13 and over)
treatment rates only about 25% instead of 75%
why might have COVID affected SSA less than other economies?
smaller drop of about 2.6% in 2020 compared to 3.5% globally
limited health impacts because of young populations
- fewer deaths per capita
large shares of rural populations where social distancing might be easier
- easier to avoid transmission
smaller share of international trade in GDP
experience from ebola with public health/lockdown/containment?
