Two stage least squares instrumental variables

[Eoghan]

I have a set of data for each state
I'm supposed to categorize: The data is from 1977-1999
VIO = violent crimes

Test the hypothesis that people in the SOUTH changed their behavior as reflected in violent crimes LOG(VIO) but only AFTER 1985 to other states
SOUTH = MD, VA, WV, NC, SC, GA, FL, KY, TN, Alabama, MISS, LA, Arkansas, and TX
AFTER = 1986 to end of data 1999

How would I specify only south states and only after 1985???

[startz]

Something like

smpl 1986 @last if state="MD" or state="VA etc.
(There may be a nicer way to do the if part)

But I have a suspicion you've misunderstood the homework.

[Eoghan]

no I haven't misunderstood the HW.

The question says to only include data from those states and 1985 <

[startz]

Well, you've read the homework and I haven't! I guess I was confused by what it means to say people "changed" their behavior. Anyhow, the smpl statement should work.

[Eoghan]

hey, I pm'd you a question, thank you for the above help!

[startz]

Please post your question here, so everyone can chime in.

[Eoghan]

The questions the professor wrote is:directions tend to be confusing

Do the instrumental variable analysis a stage at a time and also using EViews TSLS procedure
My variables are
LOG(VIO)
c
CONCEAL
INCARC_RATE
DENSITY
AVGINC
POP
PB1064
PM1029

LOG(ROB) is the variable you xhat
rob1_hat = log(rob)


How to write stages equations
Stage 1, ls x c z, fit xhat
Stage 2, ls y c xhat
Z is for all the instruments.

and then what would be my equation and instrumental variables to run TSLS?

How do I set it up using my variables?
I'm supposed to get the same coefficients in the stages and the TSLS way of testing but different std errors

[startz]

tsls y c x @ c z