You are on page 1of 6

1

Econ 3560 Applied Econometrics




HOMEWORK SET #2

(Due Friday, September 27, 2013)



Name: _________________________________________________________



INSTRUCTIONS:

Please print this sheet and staple it in front of your answer set. Homework is due at
the beginning of class. DO NOT EMAIL YOUR HOMEWORK, please hand in a
hard copy

Include the log files from your Stata homework, including required outputs (such as
histograms, scatterplots). Please highlight the relevant answers on your output,
where appropriate.

You can handwrite or type out your answers to the other (non-Stata) questions.
Please staple all your answers together.

(Do not write your answers on this sheet.)



















2



HOMEWORK SET #2


Your completed homework is due in class at the beginning of class or in my mailbox in the
Economics Department 301 Lake Hall. No credit will be given for late homework sets.


STATA

1. (More importing data, table of descriptive statistics, histogram, scatter plot, running
preliminary regressions. I suggest you use one log file for question 1, and a different
log file for question 2) The capital asset pricing model (CAPM) is an important
model in the field of finance. It explains variations in the rate of return of a stock j as
a function of the rate of return on a portfolio consisting of all publicly traded stocks
(e.g., the Dow Jones Industrial Average or the S&P 500), called the market portfolio
(m). Generally, the rate of return on any investment is measured in comparison to its
opportunity cost, which is the return on a risk-free asset (f). The resulting difference
is called the risk premium associated with that investment, because it is the reward or
punishment for making a risky investment. The CAPM says that the risk premium on
stock j (r
j
r
f
) is proportional to the risk premium on the market portfolio (r
m
r
f
).
That is:
( )
f m j f j
r r r r = |
In other words, the ratio between the two risk premia is:
( )
( )
f m
f j
j
r r
r r

= |
where r
j
is the return on stock j, r
f
is the return on the risk-free asset, and r
m
is the
return on the market portfolio.
j
| is typically called the beta value of stock j. A
stocks beta is important to investors because it reveals information about the stocks
volatility. It measures the sensitivity of stock js return to variation in the entire stock
market. As such, values of beta less than 1 indicate that the stock is defensive
because its variation is less than that of the market. A beta greater than 1 indicates an
aggressive stock, as it is very volatile. Investors usually want to know an estimate
of a stocks beta before investing in it.
The model shown in the first equation above is the economic (theoretic) CAPM
model. The econometric (empirical) CAPM model is obtained by adding an intercept

3
(
0
| ) in the economic model (even though theory says it should be zero) and an error
term (
j
c ):
( )
j f m j f j
r r r r c | | + + =
0
.
The Excel file CAPM contains data on the monthly rate of return (r
j
) of three firms
(Microsoft, General Motors, and Mobil-Exxon), the monthly rate of return on the
market portfolio (r
m
), here the S&P 500, and the rate of return on U.S. 3-month
Treasury Bills, which is the risk-free asset (r
f
). The 120 monthly observations cover
the period January 1999 to December 2008. There is a data dictionary (in pdf
format) that accompanies the dataset.
a. Import the data into Stata, properly label the variables (use whatever labels
make sense to you), and save the data as a Stata (.dta) dataset in your
working directory.
b. Use the gen command, construct the variable
f m
r r . Next, construct the
variable
f j
r r for each company in your sample. These are the two
premia discussed above.
c. Use Stata to calculate a table of descriptive statistics that effectively
summarize all the variables (except for the variable date). Focus only on the
mean, standard deviation, and range for each of the variables Then, export
these descriptive statistics from Stata to an XLS Excel file (indirectly via a
CSV file) and display them in a nicely formatted Excel table. Comment
briefly on the mean, standard deviation, and range for each of the variables.
d. Use Stata to create a histogram for the variable
f j
r r for the Microsoft
stock only. Make sure to properly label the histogram with titles, legends,
labels, notes, etc.
Comment briefly on the histogram by describing its general shape. No
statistical work is required here, I just want you to eye-ball the histogram.
(Note: Once you create a graph in Stata, the command graph export allows
you to export and convert it into a pdf document.)
e. Use OLS regression (using the robust feature in the sample .do file) to
estimate the theoretical CAPM model for each firm (separately), and comment
on their estimated beta values. Which firm appears most aggressive? Which
firm appears most defensive?
f. For the Microsoft stock only, create a scatter plot of
f j
r r (on the y-axis)
versus
f m
r r (on the x-axis). In addition, plot the fitted OLS regression
line of
f j
r r on
f m
r r within the scatter plot. Make sure to properly
label the scatter plot. You can do this with the twoway command in Stata

4
and the lfit and scatter options there is an example in the sample
program posted on Blackboard.
2. Refer back to the data set on housing prices from Homework #1.
a. Use OLS regressions (using the robust feature in the sample .do file) to
estimate the following theoretical models (each time generating the an output
for each regression):
1. ( )
i i i
livarea sprice c | | + + =
1 0

2. ( ) ( )
i i i i
age livarea sprice c | | |
2 1 0
+ + =
b. Comment on the sign and magnitude of each of the estimated coefficients for
each of the two regressions.
c. Explain why or why not the estimated coefficients differ.
d. Is there a difference in the coefficient of determination R
2
for both
regressions? Why or why not?
e. For 2(a) 1 only, generate new variables (using the gen command in Stata)
for sprice and livarea by dividing each by 1,000. Re-run the regression
results. What happens to the results relative to 2(a) 1? Explain your answer.
What happens if you divide only livarea by 1,000?



ECONOMETRIC CONCEPTS

3. Answer the following questions from the textbook:
a. Chapter 2, Question 6
b. Chapter 2, Question 9
c. Chapter 2, Question 13 (a, b, and c only)
4. (the next two questions are about expected values, refer to the rules and examples
from class notes) You are organizing an outdoor concert for next week and believe
attendance will depend on the weather. You consider the following possibilities are
appropriate:

Weather Probability = f(x) Attendance = X
Terrible weather 0.2 500
Mediocre weather 0.6 1000
Great weather 0.2 2000


5
a. Let X denote the attendance. Why is X a random variable?
b. What is the expected attendance?
c. Suppose that each ticket costs $5 and that the total cost of giving the concert is
a fixed $2,000. Let Y = profit = total sales revenue total cost = 5X 2000.
What is the expected profit?
d. If the variance of attendance is
2
X
= 240,000, find the variance of profit Y.
5. Suppose that X and Y are random variables with expected values E(X) = E(Y) =
and variances V(X)=V(Y)

=
2
.
Let Z = (X + Y)/2.
a. Find E(Z).
b. Find V(Z) assuming that X and Y are statistically independent.
c. Find V(Z) assuming that Cov(X,Y)= 0.5
2
.
6. (this question is about joint probability density functions, in class we looked at
drawing numbers on shaded or unshaded pieces of paper) C and B are two brands
of LCD TVs. During a week a certain amount of advertising appears for these
products. There may be no advertising, one form (e.g., newspaper flyers) or two
forms (e.g. newspaper flyer and a TV ad). Let C denote the level of advertising for
Company C. It can take the values C = 0, 1 or 2. Let B denote the level of advertising
of Company B; B can take the values B = 0, 1 or 2. Suppose the following table
represents the joint probability distribution of the advertising levels for these two
brands of TV.


0 1 2
0 0.05 0.05 0.05
1 0.05 0.20 0.15
2 0.05 0.25 0.15

a. What is the marginal probability distribution of Cs advertising, C? In other
words, what are the probabilities of the different values of C occurring,
without regard to the levels of B?
b. What is the expected value of C? Show your work.
c. What is the variance of C? Show your work.
d. Are the two companies advertising strategies statistically independent?
Explain.
e. Company B pays its advertising firm $5,000 per week plus $1,000 for each
level of advertising B. What is the probability distribution of Company Bs
advertising outlay, A?
B
C

6
f. What is the correlation between Company Bs advertising level (B) and its
advertising expenditure (A)? Explain.
7. (the next two questions are to get you back practicing reading normal tables) A firms
marketing manager believes that total sales X can be modeled using a normal
distribution with mean = $2.5 million and standard deviation

= $300,000. What is
the probability that the firms sales will exceed $3 million? Draw a sketch to illustrate
your calculation.
8. The length of life (in years) of a personal computer is approximately normally
distributed with mean 3.4 years and variance 1.6 years.
a. What fraction of computers will fail in the first year?
b. What fraction of computers will last 4 years or more?
c. What fraction of computers will last at least 2 years?
d. What fraction of computers will last more than 2.5 years but less than 4 years?
9. Explain what is meant by the asymptotic distribution of an estimator.
10. In class, we saw that for a simple linear regression model with only one independent
variable that the slope estimator takes a particular form:


=
.
2
_
_ _
1
) (
) )( (
x x
y y x x
i
i i
|
.

Demonstrate that this estimator for the slope parameter can be written as a linear
function of the y
i
s.
11. Explain what is meant by the term multicollinearity. Give a concrete example of this
concept (other than the example I gave in class). What are the implications for using
OLS to estimate empirical relationships?

You might also like