03. (5 points; 1 point each) Consider the following model of household savings: savings = Bo + Bi ing + B2 hhsize + B3 age + B4 educ + where inc denotes household income, hhsize is household size, age and educ denote the age and years of schooling the household head, and is an error term. a. It is often the case that household income is misreported in survey data. If we estimate savings model above using OLS, how may the presence of measurement errors in reported income affect your results? b. Ability is not controlled in the model. If you had a measure of IQ and included it as a proxy variable for ability, would it fully address the problem of not observing ability? A simple regression relating the number of crimes in university campuses to student enrollment is given by log(crime)=-6.63+1.27log(enroll)+u (1.03) (0.11) n=97, R2= .585 c. The sample for this study comes from year 1992, and many schools in 1992 did not report campus crimes. Do you think that college failure to report crimes is a problem for your estimates? Why or why not? d. The figure below provides the cumulative distribution function (cdf) of the total number of crimes. Would you trust the estimates provided in part b if you assume that failure of some colleges to report crimes is NOT a problem? Why or Why not? 1 calf 2 0 500 1000 total campus crines 1500 2000 e. Is there any reason why you would like to estimate this model using Least Absolute Deviations method? Explain. 03. (5 points; 1 point each) Consider the following model of household savings: savings = Bo + Bi ing + B2 hhsize + B3 age + B4 educ + where inc denotes household income, hhsize is household size, age and educ denote the age and years of schooling the household head, and is an error term. a. It is often the case that household income is misreported in survey data. If we estimate savings model above using OLS, how may the presence of measurement errors in reported income affect your results? b. Ability is not controlled in the model. If you had a measure of IQ and included it as a proxy variable for ability, would it fully address the problem of not observing ability? A simple regression relating the number of crimes in university campuses to student enrollment is given by log(crime)=-6.63+1.27log(enroll)+u (1.03) (0.11) n=97, R2= .585 c. The sample for this study comes from year 1992, and many schools in 1992 did not report campus crimes. Do you think that college failure to report crimes is a problem for your estimates? Why or why not? d. The figure below provides the cumulative distribution function (cdf) of the total number of crimes. Would you trust the estimates provided in part b if you assume that failure of some colleges to report crimes is NOT a problem? Why or Why not? 1 calf 2 0 500 1000 total campus crines 1500 2000 e. Is there any reason why you would like to estimate this model using Least Absolute Deviations method? Explain