Lbymet Pset1
Lbymet Pset1
Lbymet Pset1
Economics Department
LBYMETF PSET1
Directions: Implement all the tasks using Stata. Email your .do files on or before Feb. 13,
04:14pm. Comment the interpretations and explanations in the .do file as well (with an asterisk
(*)).
1. Why is the Linear Probability Model better than Ordinary Least Squares when estimating a
dependent variable that only takes on a value of 0 or 1?
2. Use http://www.ats.ucla.edu/stat/stata/dae/binary.dta. Regress admit with gre, gpa, and rank. If
you proceed to predict Yhat after the regression and list the values, why are some of the Yhat
values negative?
3. Use this data from the internet. http://www.ats.ucla.edu/stat/stata/dae/binary.dta
a. Make a logit regression using admit gre gpa rank.
b. If you belong to a good school ie. rank=1, what are the chances that you will be
accepted?
c. If your GPA is 4.0, you belong to a good school ie. rank=1, and you have a GRE score
of 800, what are the chances of you being accepted?
4. What is the difference between logit and probit?
5. Mortgage lenders are interested in determining borrower and loan factors that may lead to
delinquency or foreclosure. In the file lasvegas.dat are 1000 observations on mortgages for
single-family homes in Las Vegas, Nevada, during 2008. The variable of interest is
DELINQUENT, an indicator variable = 1 if the borrower missed at least three payments (90 or
more days late), but zero otherwise. Explanatory variables are LVR = the ratio of the loan
amount to the value of the property; REF = 1 if purpose of the loan was a ‘‘refinance’’ and = 0 if
loan was for a purchase; INSUR = 1 if mortgage carries mortgage insurance, zero otherwise;
RATE = initial interest rate of the mortgage; AMOUNT = dollar value of mortgage (in
$100,000); CREDIT = credit score, TERM = number of years between disbursement of the loan
and the date it is expected to be fully repaid, ARM = 1 if mortgage has an adjustable rate, and =
0 if mortgage has a fixed rate.
(a) Estimate the linear probability (regression) model explaining DELINQUENT as a
function of the remaining variables. Are the signs of the estimated coefficients
reasonable?
(b) Interpret the coefficient of INSUR. If CREDIT increases by 50 points, what is the
estimated effect on the probability of a delinquent loan?
(c) Compute the predicted value of DELINQENT for the final (1000th) observation.
Interpret this value.
(d) Compute the predicted value of DELINQUENT for all 1000 observations. How many
were less than zero? How many were greater than 1? Explain why such predictions are
problematic.