Smith School of Business
MGMT 880/990
Winter 2018
Professor Olena Ivus
Final Exam Practice Questions
1. Consider the difference-in-means estimator ¯
d= ¯y1−¯y0, where ¯y1is the sample
average of yiwith wi= 1 and ¯y0is the sample average of yiwith wi= 0.
(a) You would like to estimate the average treatment effect on the treated:
T T =E(y1−y0|w= 1)
Show that ¯
dis a biased estimator of the average treatment effect, and the bias
is given by E(y0|w= 1) −E(y0|w= 0).
(b) Let y0be the earnings someone would earn in the absence of job training,
and let w= 1 denote the job training indicator. Explain the meaning of
E(y0|w= 1) < E(y0|w= 0). Intuitively, does it make sense that E(¯
d)< T T ?
2. The table in Attachment A is from Bernard, Redding, and Schott (2010) “Multiple-
Product Firms and Product Switching,” American Economic Review, 100:1, 7097.
In column 2 (Multiple product), the characteristics of single- and multiple-product
firms are compared for the 1997 census. The column reports the results from OLS
regressions of form characteristic (in logs) on a dummy variable for the multiple-
product firms (one if a firm produces multiple products) and main industry fixed
effects.
(a) Interpret the coefficient 0.58.
(b) What variation is used to estimate this coefficient?
(c) Why did the authors cluster standard errors by main industry?
3. The table in Attachment B is from Neumark (2002) “Youth Labor Markets in the
United States: Shopping Around vs. Staying Put” The Review of Economics and
Statistics. Neumark is interested in determining whether having multiple jobs as a
young person hurts or helps labor market outcomes later in life. The equation of
interest is:
ln(wi) = Xiβ+αSi+ui
where iindexes individuals, wis the adult wage, Xtis a vector of exogenous covari-
ates, and Sis longest tenure obtained (in years) on a job in the five years following
schooling (measures early job stability).
1
