Econ Inequality Midterm 1

Between vs within

between - inequality acrosss groups (gender, racial, college wage gaps)

within: inequality in a given group (ex. inequality within women)

total inequality = between + within

outcome vs opportunity

outcomes: look at this mainly because it is easier to measure

opportunity: does the starting point matter

according to K+M, what happened to the college wage premium

decreased in 70s because supply grew quickly and increased in 80s because supply growth slowed and demand started to increase for college workers

incerase in returns to experience

decrease in gender wage gap

what is the college wage premium

wh/wl - college workers wage relative to high school workers

why does wage equal the marginal product in KM?

firms pay each labor type its marginal product

what determines the college wagre premium

relative demand/productivity (AH/AL) and relative supply (H/L)

what happens to the wage premium when (H/L) rises?

holding all else constant, it would fall (more college workers means more competition and lower wages)

what happens to the college wage premiun when AH/AL rises?

rises b/c demand/rpdocutvity of college workers has increased

how is unobserved relative demand modeled in KM?

ln(AH/AL) = Bo0 + B1t

says demand for college workers has been increasing at roughly a constant rate over time

KM estimating equation? (with relative supply and relative demand observable)

ln(wh/wl) = Bo0+ B1t - (1-sigma)ln(H/L) + error term

relative demand - relative supply

what does sigma represent in KM?

elasticity of substituion between college and high school labor

larger sigma = more substitutable = smaller effect of supply changes on wages

coefficient of relative supply in KM

-1/sigma

how to get beta from sigma in KM to determine

the (-1/sigma) that is * relative supply is your coefficient

so if you run regression and get a # for Beta - set it equal to -1/sigma to solve for sigma

main takeaway of autor automation paper

number of jobs may increase or not chance but the types of jobs does change

according to automation paper why are there still so many jobs

jobs that complement machines increase productivity and lower the cost of production so there is more ability to expand and the demand for these jobs increases

2 questions to classify jobs according to automation paper

1. can it be codified into explicit rules?

2. does a machine replace a worker of make a worker more productive?

examples of routine cognitive and routine manual

cognitive: wor prodcessing

manual: repetitve assembly, machine tending

exmaples of non routine cognitive and manual

cognitive: doctor, lawyer, problem sovling

manual: waitress, cooks

anything non routine is what cannot be replaces by automation: high skill and low skill jobs

polanyis paradox

much of human knowledge/competence is facit + tactile

ex. recognizing a few, juding a situation

no explicit rules so cant be codified

this is why non routine jobs have not been taken by automation (ex. cant have a robot waitress)

O-Ring

production is modeled as a chain of tasks- each relying on qulity of previous task

output depends on weakest link

if some tasks are automated, this can improve production of those or decrease cost so there is an increase in marginal benefit of tasks done by others, more skilled humans increased

less routine tasks become more valuable and profotable + more routine replaced by machines

ex. bank teller job evolved into more personal connection while routine aspects were taken

why according to automation paper is there less evidence for wage polarization

because middle skill people moved into low skill jobs so their wages did not increase since there was a larger supply of workers

only high skill workers wages increase because demand for them went up and there is more of a barrier of entry since you need more schooling

compartive advantage

countries specilaize in making the good that requires the facotr of production they have the most abundantly

for trade paper, why cant we just regress US employment on US imports from china?

need to control for stuff, different parts of US affected differntly

endogeneity: an increase in employment could be cuased by an increase in imports

how to autor dorn and hanson address the issue of endogeneity?

1. use commuting zones: local labor markets as a unit of observation

   1. manufacturing industries clutter in certain places and differnt locations specialize in diff things

2. construct a shift-share measure local CZ exposure is builf by looking at pre-shock industry mix * national industry level imports growth

   1. import exposure in a CZ will add up the industry exposure for that CZ across all industries

   2. pre period weights are used to understand/fix the mix of industries in an area, this allowes us to understand how exposed the labor mkt is to increase in imports

3. use an instrument to eliminate concerns about endogeneity

   1. use an import growth in other developed countries to predict US import growth by industry

China paper findings

1. decerease in manufacturing employment

2. overall employment/pop down, LF participation down, in all industries unemployment up

3. wages decreased, concetrated in non college workers

4. disability, unemployment payments, use of medicaid and social secuity all up

5. adjustment in local labor markets is slow and weak (few people leave/retrain)

imports account for about 25% for the decrease in aggregate US employment

key takeway from balu and khan

uncontrolled gap is not less meaningful than the controlled one

if equal pay for equal work holds, a remaining uncontrolled gap means that woen are underrepresented in high paying jobs

this is also the idea that sorting is an outcome and why we’ll interpret occupation and industry controls with care

what happened to the gender wage gap in the 1980s

big convergence

what was the primary driver of the convergence of the convergence of the gender wage gap

education and experience gap fell

women caught up in terms of years of schooling, now more educated

men contrinue to have more eperience and managerial roles, but women caught up mostly

collective bargaining decreased for men so women now have more (women are teachers and manufacturing jobs went away)

what explains a large portion of the gender wage gap in 2010

occupation and industry

even though women have upgraded their occupations, return to male-dominated occupations has increased by more

2 regressions for oaxaca blinder decomp

male: ln(wi,m) = Xi,mBm + ui,m

B0m+b1mXi,m + ui,m

fem: ln(wi,f) = Xi,fBf + ui,f

B0f + B1fXi,f + ui,f

Xi,m and Xi,f are vectors of charcteristics like education experience

what is the raw gap in oaxanca for male and female

bar ln(wm) - bar ln(wf)

these are the averages wages for male and female

bar ln(wm) = Bhat0m + B1hatXm

bar ln(wf) = Bhat0f + B1hatXf

X bar (all of these Xs) is average observation for X

what is the counterfactual for the Oaxaca

what would a woman be paid if she keot her own characteristic but was paid on the male wage schedule

B0m hat + B1m hat + Xbarf

(from the regression if u only include the male wage part and then X are characteristics)

rewritten raw gap with counterfactuals

rewritten gives us the explained gap

ln(wm) - ln(wf) + counterfactual - counterfactual

=

explained gap in oaxaca

B1hatm(Xmbar-Xfbar)

>0 men have more of something (ex. experience)

<0 women have more of something

unexplained gap in oaxaca

Xfbar(Bm-Bf)

raw gap - explained

what does Xmbar = Xfbar

and Bmhat = bfhat mean

first is when men and women have the same characteristics

when men and women are being paid same for every characteristic

in our code model (not blau and khan) what happened to the explained gao by 2012?

went under 0, unexplained becomes larger than the raw gap

how much of the gap to blau and khan explain

62%

what does actual vs potential experience measure for the gender wage gap

potential experience misses career gaps, especially for women

does a larger unexplained gap in gender wage gap mean discrimination?

could be, but could also be other omitted variables such as

• Productivity - hours worked

• Tenure

• Actual experience

• More detail on actual tasks in the job - more skill intensive jobs could pay more

why restrict workers to full time in oaxaca

creates more comparable sample of workers strongly attched to the workforce

downside - women may differ significantly from women not full time

this is why selection bias is greater for women - smaller more selected share of women is working

with selection of women into workforce, if only high-productive women select in, what happens to wage gap?

observed wage gap is too small because women’s wage is observed too high

he women we observe working have higher average wages than the full population of women would have → this makes women look closer to men's wages → therefore, the observed gender wage gap is artificially small.

what has happened with the female LF participation rate

was rising then plateaud around 57% in 2013