|
Research
My research
focuses mainly in Development and Labor Economics. My research encompasses
three different countries, namely India,
Kenya and the USA. The
projects are described below.
India:
Moving
Women: Household Composition, Labor Demand and Crop Choice:
This paper estimates the effect of increases in the
relative value of female labor induced by rainfall shocks on the gender
composition of rural agricultural households in India. Much of the previous
literature in this vein has focused exclusively on the gender composition of
young children due to excess female mortality. Using a unique panel dataset
from India,
I focus instead on the changes in the gender composition of prime aged
adults. Rice production is known to be more intensive in female labor than
wheat production, and is more dependent on rainfall. I compare how rice and wheat households
adjust the composition of their adult membership in response to rainfall
shocks that differentially raise the relative marginal products of female
labor, controlling for farm and household-level heterogeneity. Consistent
with a model of household composition and crop choice in the presence of
imperfect labor markets, I find that positive rainfall shocks are associated
with an increase in the relative number of prime age females in rice
households. Timing of marriage of
daughter’s of household heads is one of the main mechanisms adjusting the
gender composition of prime aged adults in these households. I find that a
one standard deviation increase in rainfall decreases the rate of marriage
among young adult females in rice households by approximately 10 percent
relative to wheat households. Dowries paid out by rice households also
decline by a similar magnitude, indicating a rise in the value of female
labor.
USA:
An Empirical
Analysis of Black-White Employment Differences over the Business Cycle
(Joint with Yusuf Soner Baskaya)
Previous research on the US labor market has shown that
unemployment rates of blacks have not only been substantially higher than
that of whites over the last four decades, but also these differences have
been amplified during recessions. However, the extent to which these
differences reflect unobserved skill and productivity or other factors such
as discrimination remains a matter of some debate. We propose the use of wages earned in the
previous year as a measure of a worker’s skill and productivity. Using the
Current Population Survey March Supplement and the National Longitudinal
Survey of Youth, in each year from 1976 to 2003, we compare the employment
outcomes of black and white workers who earned the same wage in the previous
year. After adjusting for worker skill
and productivity in this way, we find that the unexplained gap in
unemployment between blacks and white falls slightly, consistent with the
view that some of the difference arises from unmeasured skill and
productivity difference between the races. Nevertheless, the gap remains
large and significant, and the pattern of larger gaps during recessions is
unaffected. Consistent with previous work, we find that these differences do
not exist for individuals in the highest part of the skill distribution. Our
results improve upon the existing literature on the sources of black-white
differences in employment outcomes by mitigating the problems related to
unobserved individual productivity and skills. Under general assumptions
about labor supply changes over the business cycle, supported by behavior of
reported reservation wages of black and white youth, these results may
indicate discrimination against blacks.
Kenya:
School
Quality and Student Achievement in Kenya: A Regression Discontinuity
Approach
I utilize data from the Kenyan secondary school
system to obtain causal estimates of the effects of school quality on student
achievement. Whereas most studies on the effects of school quality on student
achievement generally face difficulties in obtaining unbiased estimates due
to the non-random selection of students into schools, the placement of
students into government secondary schools in Kenya is based solely on
national primary schools test scores and district quotas. I utilize the
random variation induced by this quota system to isolate the treatment effect
of school quality on subsequent student performance in the national high
school examination. Using a unique data set containing high school test
scores, primary school test scores, district of origin and school level
information for every high school exam taker in the country, I compare the
high school examination outcomes of students from the same district who had
very similar primary school test scores but were assigned to different
schools due to the quota system. I extend the analysis further to examine
whether peer effects or school inputs have greater effect on student
achievement in secondary schools in Kenya..inconvienence.inconvienence.inconvienence.
|
