Alocal government that wants to tackle unemployment in its village has started
a program by which it guarantees 100 days per year of employment to those that
enroll. with this program, the local government seeks to increase the employability of
participants, and thus its final objective is to increase the total number of employment
days per year (both within and outside the program) for those that participate.
enrollment to the program is open to the whole village, and 50% of the village
decides to join.
(a) what are the treatment (wi) and outcome (yi) variables?
(b) what would the conterfactual be? explain in words and using notation that
was introduced during lecture.
(c) could selection bias be a problem here? if yes, explain why (both in words and
mathematically).
(d) considering your argument in question (c) above, what is your prediction about
the sign of the selection bias in this case? would a naive assessment of the
observed differences under- or over-estimate the causal effect of treatment?
(e) what would have been different if the local government had offered the pro-
gram to a randomly selected 50% of the population? explain in words and mathematically.