ECON423 - Exam 2
probit(h) sharet c trated # autos by truck=
# autos by truck=prob truck%*#total auto shipped
df for F statistic
# restrictions and n-k
LRTS (likelihood ratio test statistic)
-2(Lr-Lu)=-2(-210--205)=-2(5)=-10
firms.est(h) cost? @ c q? # of independent period fixed effects
0 because (xxx) does not contain per=f
Phi(-2.58) Phi(2.58)
0.5% 99.5%
# of slope coefficients for bivariate
1
firms.est(h,cx=f,per=f) cost? @ q? # of intercepts
1
15 observations of 10 states dimensions of u
150 (15x10) by 1
15 observations of 10 states dimensions of cov(u)
150 x 150
z=-1 z=1
16% 84%
Phi(-1.96) Phi(1.96)
2.5% 97.5%
firms.est(h,cx=f,per=f) cost? @ q? # of independent cross section fixed effects
3 (4 firms - 1)
firms.est(h,cx=f,per=f) cost? @ q? # of slope coefficients
4 (one for each firm)
Phi(-1.645) Phi(1.645)
5% 95%
Phi(0)
50%
truck/rail price difference of zero, # autos shipped by truck=
50% of total because z=0 => Phi(0)=50%
firms.est(xxx) COST? Q? yrII yrIII @ c geometry
6 parallel lines
firms.est(h,cx=f,per=f) cost? @ q? # of independent period fixed effects
9 (10 years - 1)
EViews log likelihood
@Logl
wald statement for truck/rail price difference of zero
@cnorm(c(1))=0.5
ls(h) approved c white female female/male difference in white/black difference = 0% coincidence? why or why not?
NO Because coefficient of interacted dummy variable=0 NB white/black issue has been fixed here
trated
Truck rate - Rail rate
xxxx.est(h) Y? X? @ C label c(i)
X? c(1) C c(2), c(3)
probit(h) call_back c female EViews expression to test Ho that female-male difference is zero
[@cnorm(c(1)+c(2))]-[@cnorm(c(1))]=0
fixed effects estimated intercept
average of separate estimated intercepts
firms.est(xxx) COST? Q? yrII yrIII @ c C (intercept)
average of the averages in each row
LRTS distribution
chi squared
we associate @covariance with
coefficient covariance matrix
cx=f
cross section fixed effects sum to zero
firms.est(xxx) COST? Q? yrII yrIII @ c c(4) =
firm A @ yrI (table)
firms.est(xxx) COST? Q? yrII yrIII @ c c(5) =
firm B @ yrI (table)
fixed effects c--XXX
fixed effects estimated intercept+c--XXX=separate estimated intercepts
Estimated intercept for cx=f,per=f
intercept+cross fixed effects+period fixed effects
firms.est(xxx) COST? Q? yrII yrIII @ c c(4) is the _____ for firm ____ in year ____
intercept, A, 1
countries.est(h) yb? b? g? @ c int_us=48 int_can=48 countries.est(h, cx=f) yb? b? g? intercept reported on EViews c--US c--CAN
intercept=48 c--US=0 c--CAN=0
15 observations of 10 states dimensions of @COVARIANCE
k x k
firms.est(h,cx=f,per=f) cost? @ q? k=?
k=(# of intercepts)+(# of slope coefficients)+(# of independent cross section fixed effects)+(# of independent period fixed effects)=17
k
number of regressors
per=f
period fixed effects sum to zero
probit(h) sharet c trated prob truck%=
prob truck%=phi(z_hat)
firms.est(xxx) COST? Q? yrII yrIII @ c is c(1) the marginal cost?
yes
probit(h) call_back c female c(1)=a1 c(2)=a2 ahat1+ahat2=0 (y,n)
yes
firms.est(xxx) COST? Q? yrII yrIII @ c c(2) =
yrII-yrI
firms.est(xxx) COST? Q? yrII yrIII @ c c(3) =
yrIII-yrI
probit(h) sharet c trated z_hat=
z_hat=c(1)+c(2)*trated
probit(h) sharet c trated ΔP=
ΔP=Δtrated
arc-slope
ΔQ/ΔP
