564 Panel Data Methods
problems of retransformation back to the “natural scale” in the presence of individ-
ual heterogeneity (Manning, 2006). The RAND data is available over the internet
and has been used to test more recent developments of econometric methods
(Bago d’Uva, 2006; Deb and Trivedi, 2002; Gilleskie and Mroz, 2004; Vera-
Hernandez, 2003).
More recently, randomized experiments have begun to play an influential role
in research and policy in developing countries. This is exemplified by the stud-
ies of Gertler (2004) and Miguel and Kremer (2004). The Mexican government’s
PROGRESA program, which was initiated in 1997, has received considerable atten-
tion and has influenced policy throughout Latin America. The program relies on
conditional cash transfers that are designed to influence the use of health and
welfare services for children in poor families. It covers 2.6 million families in
50,000 rural villages. The program focuses on health, hygiene and nutrition. It
links substantial cash transfers, on average amounting to 20–30% of household
income, to the use of prenatal care, well-baby care and immunization, nutri-
tion monitoring and supplementation, preventive check-ups and participation
in educational programs. PROGRESA works by first selecting whole communities
to participate in the scheme and then selecting households within those com-
munities that satisfy the eligibility criteria to receive the benefits of the scheme.
Financial constraints on the implementation of PROGRESA meant that its intro-
duction was phased. To make the implementation equitable, communities were
selected randomly to receive the benefits either immediately or with a delay. The
random phasing provides researchers with an ideal opportunity to use a random-
ized design in the evaluation of the impact of the program. Of the communities
selected for the program, 320 were randomly selected to receive the intervention in
August–September 1998 with the remaining 185 delayed for two years. The com-
munities in the control group were not informed that they would eventually receive
the program, reducing the scope for anticipation of treatment to influence the
outcome.
Gertler (2004) focuses on health outcomes among children. These include self-
reported morbidity, measured by illnesses in the past month, as reported by the
child’s mother, and objective measures including anthropometric measures of
height and stunting and a biomarker for anaemia (haemoglobin levels). The anal-
ysis is restricted to those households, in both the treatment and control groups,
that satisfy the eligibility criteria for PROGRESA. Although the data is random-
ized, multivariate regression models are used to control for observed covariates
and to reduce idiosyncratic variation. Individual and village random effects are
included, the latter to allow for the clustered sampling. The results show signifi-
cant improvements in both self-reported and objective measures of health and the
impact increases with length of exposure to the program. Gertler (2004) is careful to
note that the comparison of treated and controls does not explain the mechanism
behind this effect: for example, it is not possible to say whether an unconditional
transfer would have had the same effect as the conditional one.
In their “worms” paper, Miguel and Kremer (2004) analyze a randomized exper-
iment to evaluate the impact of the Kenyan Primary School Deworming Project