19.7 In his article, “Instrumental-Variable Estimation of Count Data Models: Applications to Models of Cigarette Smoking Behavior”, Review of Economics and Statistics (1997, pp. 586–93), John Mullahy wanted to fi nd out if a mother’s smoking during pregnancy adversely aff ected her baby’s birth weight. To answer this question he considered several variables, such as natural log of birth weight, gender (1 if the baby is male), parity (number of children the woman has borne), the number of cigarettes the mother smoked during pregnancy, family income, father’s education, and mother’s education. Th e raw data can be found on the website of Michael Murray (http://www. aw-bc.com/murray/). Download this data set and develop your own model of the eff ect of mother’s smoking during pregnancy on the baby’s birth weight and compare your results with those of John Mullahy. State your reasons why you think that a standard logit or probit model is suffi cient without resorting to IV estimation.