BEGIN:VCALENDAR
VERSION:2.0
PRODID:talks.ox.ac.uk
BEGIN:VEVENT
SUMMARY:Measurement Error without Exclusion: the Returns to College Select
ivity and Characteristics - Karim Chalak (University of Virginia)
DTSTART;VALUE=DATE-TIME:20180504T131500Z
DTEND;VALUE=DATE-TIME:20180504T143000Z
UID:https://talks.ox.ac.uk/talks/id/58022c54-a58a-4080-8253-0e6996063f36/
DESCRIPTION:Abstract\n\nThis paper studies the identification of the coeff
icients in a linear equation when data on the outcome\, covariates\, and a
n error-laden proxy for a latent variable are available. We maintain the c
lassical error-in-variables assumptions and relax the assumption that the
proxy is excluded from the outcome equation. This enables the proxy to dir
ectly impact the outcome and allows for differential measurement error. Wi
thout the exclusion restriction\, we first show that the coefficients on t
he latent variable\, the proxy\, and the covariates are not identified. Th
en\, we derive the sharp identification regions for these coefficients und
er either or both of two auxiliary restrictions. The first restriction wea
kens the assumption of “no measurement error” by imposing an upper bou
nd on the net of the covariates “noise to signal” ratio\, i.e. the rat
io of the variance of the measurement error to the variance of the latent
variable given the covariates. The second restriction weakens the proxy ex
clusion restriction by specifying whether the latent variable and its prox
y affect the outcome in the same or the opposite direction\, if at all. Us
ing the College Scorecard data\, we employ this framework to study the fin
ancial returns to college selectivity and characteristics. Here\, college
selectivity\, defined as the average SAT score of a student cohort\, serve
s as a proxy for the latent average scholastic ability and is included in
the average earnings equation. We obtain an informative upper bound on the
return to college selectivity which becomes smaller upon conditioning on
the instructional expenditures per student and the completion rate. Furthe
r\, we obtain tight bounds on the returns to the college characteristics a
nd find that conditioning on the composition of majors reduces the magnitu
de of the bounds on the effect of some of these characteristics\, such as
the gender composition.\n\nLink to paper:\n\nhttp://people.virginia.edu/~k
mc2yf/KC%2005-21-17_Exclusion_Identification_G.pdf\n\nSpeakers:\nKarim Cha
lak (University of Virginia)
LOCATION:Manor Road Building (Seminar Room C)\, Manor Road OX1 3UQ
URL:https://talks.ox.ac.uk/talks/id/58022c54-a58a-4080-8253-0e6996063f36/
BEGIN:VALARM
ACTION:display
DESCRIPTION:Talk:Measurement Error without Exclusion: the Returns to Colle
ge Selectivity and Characteristics - Karim Chalak (University of Virginia)
TRIGGER:-PT1H
END:VALARM
END:VEVENT
END:VCALENDAR