In: Statistics and Probability
Consider the following model to estimate the effects of several variables, including cigarette smoking, on the weight of newborns:
log(bwght) = β0 + β1male + β2parity + β3log(faminc) + β4packs + u
where male is a binary indicator equal to one if the child is male; parity is the birth order of this child; faminc is family income; and packs is the average number of packs of cigarettes smoked per day during pregnancy.
(i) Why might you expect packs to be endogenous?
( ) There may be observed health factors uncorrelated with smoking behavior that affect infant birth weight.
( ) There may be unobserved health factors correlated with smoking behavior that affect infant birth weight.
( ) There may be observed health factors correlated with smoking behavior that affect infant birth weight.
( ) There may be unobserved health factors uncorrelated with smoking behavior that affect infant birth weight.
(ii) Use the data in BWGHT.dta to estimate the structural equation. First, use OLS. Then, use 2SLS, where cigprice is an instrument for packs.
(iii) Estimate the reduced form for packs. The reduced form estimates show that cigprice does not significantly affect packs is _________ (TRUE or FALSE)
(iv) What do you conclude about identification of the structural equation using cigprice as an instrument for packs?
( ) Variable cigprice fails as an IV for packs because cigprice is partially correlated with packs.
( ) Variable cigprice does not fail as an IV for packs because cigprice is partially correlated with packs.
( ) Variable cigprice fails as an IV for packs because cigprice is not partially correlated with packs.
( ) Variable cigprice does not fail as an IV for packs because cigprice is not partially correlated with packs.
(i) There may be unobserved health factors correlated with smoking behavior that affect infant birth weight.
There may be women smokers who may also be involved in other activities like drinking coffee or eating less nutritious food or consume alcohol which can impact on infant birth weight.
(ii) The difference between OLS and 2SLS in the estimated effect of packs on bwght is huge. ANSWER: TRUE
With the OLS estimate, one more pack of cigarettes is estimated to reduce bwght by about 8.4 % and IS statistically significant.
The 2SLS has the OPPOSITE sign, is HUGE in magnitude, and IS NOT statistically significant.
(iii) The reduced form estimates show that cigprice does not significantly affect packs is TRUE.
(iv) Variable cigprice fails as an IV for packs because cigprice is not partially correlated with packs.