Class Notes (1,100,000)
CA (620,000)
McGill (30,000)
POLI (3,000)
Lecture 10

POLI 441 Lecture Notes - Lecture 10: Standard Deviation, Observational Equivalence, Trae Tha Truth

Political Science
Course Code
POLI 441
Krzysztof Pelc

of 1
Coding individuals
Scheve and Slaughter then code skill level using 2 variables:
occupation wages
education years
then they code trade vulnerability in 2 ways:
net export share (industry level exports (-) imports)
tariffs: built-in assumption about comparative disadvantage and demands for
theoretical expectations (over support for protection)—HYPOTHESES
occupation wage and education years are negatively related to approval of protection
sector tariffs make you more like to want tariffs and greater protection
sector net export share make you less pro-barriers
Scheve and Slaughter’s results
occupation wage: strongly supported
education years: strongly supported
sector tariffs: supported, but
sector net export share: supported, but
implications: there is relatively high intersectoral labour mobility in the US over time
horizons relevant to individuals evaluating trade policy (HO dominates RV)
another testable expectation
trade policy is one of the forces affecting the level of regional economic activity
freer trae tends to shrink comparative-disadvantage industries and expand
comparative-advantage ones
regions with a higher concentration of activity in sectors with a comparative
disadvantage are more vulnerable to adverse housing-demand shocks from
think “Roger and Me"
even individuals working in non-tradables in Flint, Michigan, were affected by GM’s
plant closing down, and its jobs shipped abroad
if you were an individual working in non-tradable in Flint, you don’t care (in theory)
about liberalization, but you ACTUALLY DO
the value of houses decreased
this is one channel through which people take on the preferences of their
neighbours if they own their house
take on regional preferences, regardless of where you stand on trade
results of Housing Effect
Scheve and Slaughter find strong support for their expectations: an interaction effect
for homeowners, an increase in county trade exposure from its sample mean to one
standard deviation above its mean increases the odds of being ag. trade by up to
regional/town vulnerability and home ownership are the variables
an interaction test: homeowners and county exposure, by themselves, insufficient
implications this holds for “homo economics” assumptions?
a closer look at education
Hainmueller and Hiscox problematize findings supporting the HO model’s expectations
for the effect of education on preferences
they suggest an observationally equivalent story
is the link b/w education and trade preferences a question of distributional effects, or
exposure to certain ideas?
trade preferences and “ideas"
two candidates for the role of ideas on trade:
familiarity with economic concepts (along the lines of intellectual history)
cultural factor: greater openness to cosmopolitanism translating into foreign
policy ideas
H & H use the same survey as S & S
but how do they distinguish between
education as ideas
education as distributional effects?
they look at retired individuals
they also claim that we should see linear effects for education if its role is purely
distributional, and non-linear effects for an ideas role