applied econometrics
Instructions: The coursework consists of a computer exercise (100 marks). You have to
submit ONE PDF le where you write up your answers, and present all your results and
computer code in a legible and intelligible fashion. You will lose marks if tables are
misaligned, graphs not displayed, etc. The deadline for submitting your course-
work is November 21, 2017, at 4pm. Late submissions up to 24 hours will be
capped. Submissions which are late more than 24 hours are a fail. One last
piece of advice: There is almost no question (regarding the use of Stata) which cannot be
answered by a quick google search…
1 Computer Exercise – 100 marks
The question we will be addressing in this exercise is whether or not trade liberalisation leads
to income convergence or divergence among the liberalising countries. We will be looking
at two measures. The rst is the standard deviation of the natural logarithm (henceforth
denoted log) of real per worker income y (y = log(real per worker income)), which we will
denote by . If for a group of countries this standard deviation decreases, it means that
their per worker incomes converge (if it were zero, all countries would have the same per
worker income). You will estimate:
t = 0 + 1Lt + 2t + 3t Lt + ut (1)
where Lt is a dummy equal to one for all t after the trade liberalization, and t is year. For
convenience, for this exercise you should let t start at t = 1.
The second measure will be constructed in the following way: Let yt be the average log per
worker income over all countries in year t. Country i’s distance to the average in year t is
zi;t = yit
