You are on page 1of 8

Problem1:

The demand for roses was estimated using quarterly figures for the period 1971 (3rd quarter) to
1975 (2nd quarter). Two models were estimated and the following results were obtained:

Y = Quantity of roses sold (dozens)


X2 = Average wholesale price of roses ($ per dozen)
X3 = Average wholesale price of carnations ($ per dozen)
X4 = Average weekly family disposable income ($ per week)
X5 = Time (1971.3 = 1 and 1975.2 = 16)
ln = natural logarithm
The standard errors are given in parentheses.
A. ln Yt ^ = 0.627 - 1.273 ln X2t + 0.937 ln X3t + 1.713 ln X4t - 0.182ln X5t
(0.327) (0.659) (1.201) (0.128)
R2 = 77.8% D.W. = 1.78 N = 16

B. ln Yt = 10.462 - 1.39 ln X2t


(0.307)
R2 = 59.5% D.W. = 1.495 N = 16

Correlation matrix :
ln X2 ln X3 ln X4 ln X5
ln X2 1.0000 -.7219 .316 -.7792
0
ln X3 -.7219 1.0000 -.1716 .5521
ln X4 .3160 -.1716 1.0000 -.6765
ln X5 -.7792 .5521 -.6765 1.0000

a) How would you interpret the coefficients of ln X2, ln X3 and ln X4 in model A?


What sign would you expect these coefficients to have? Do the results concur with your
expectation?
b) Are these coefficients statistically significant?
c) Use the results of Model A to test the following hypotheses:
i) The demand for roses is price elastic (so sánh vs 1)
ii) Carnations are substitute goods for roses (correlation = -.7219)
iii) Roses are a luxury good (demand increases more than proportionally as income rises)
d) Are the results of (b) and (c) in accordance with your expectations? If any of the tests are
statistically insignificant, give a suggestion as to what may be the reason.
e) Do you detect the presence of multicollinearity in the data? Explain.
f) Do you detect the presence of serial correlation? Explain
g) Do the variables X3, X4 and X5 contribute significantly to the analysis? Test the joint significance
of these variables. (t-test cho từng cái)
h) Starting from model B, assuming that at the time point of January 1973, there was a disaster that
heavily affected the quantity of roses produced. Suggest a model to check if we have to use two
different models for the data before and after the disaster. (Using dummy variable).
Problem 2:
Two large US corporations, General Electric and Westinghouse, compete with each other and
produce many similar products. In order to investigate whether they have similar investment
strategies, we estimate the following model using pooled time series data for the period 1935 to
1954 for the two firms:

INVt = 1 + 2DVt + 3Vt + 4DV*Vt + 5Kt + 6DV*Kt + ut (1)

where INV = gross investment in plant and equipment


V = value of the firm = value of common and preferred stock
K = stock of capital
DV = 0 if General Electric (observations 1 to 20)
= 1 if Westinghouse (observations 21 to 40)

All three continuous variables are measured in millions of 1947 dollars. Pooling the data yields
40 observations with which to estimate the parameters of the investment function. However,
pooling is valid only if the regression parameters are the same for both firms. In order to test this
hypothesis, intercept and slope dummy variables are included in the model.

Dependent Variable: INV


Method: Least Squares

Sample: 1 40
Included observations: 40
Variable Coefficient Std. Error t-Statistic Prob.
C -9.956306 23.62636 -0.421407 0.6761
DV 9.446916 28.80535 0.327957 0.7450
V 0.026551 0.011722 2.265064 0.0300
DV*V 0.026343 0.034353 0.766838 0.4485
K 0.151694 0.019356 7.836865 0.0000
DV*K -0.059287 0.116946 -0.506962 0.6155
R-squared 0.827840 Mean dependent var 72.59075
Adjusted R-squared 0.802523 S.D. dependent var 47.24981
S.E. of regression 20.99707 Akaike info criterion 9.064124
Sum squared resid 14989.82 Schwarz criterion 9.317456
Log likelihood -175.2825 F-statistic 32.69818
Durbin-Watson stat 1.121571 Prob(F-statistic) 0.000000

(a) Interpret all the coefficient estimates, stating whether the signs are as you would expect, and
comment on the statistical significance of the individual coefficients.
(b) Comment on the overall fit and statistical significance of the model.
(c) The Jarque-Bera statistic is 7.77 and its p-value is 0.02. What can you conclude about the
distribution of the disturbance term? Why is this test important?
(d) On the basis of the above results, is pooling the data from the two firms appropriate? Explain.
(e) An alternative way of testing whether pooling the data is appropriate, without using dummy
variables, is to use the Chow breakpoint test. Referring to table below, briefly discuss how
the test works and whether the results are consistent with the earlier model (which includes
dummy variables).

Chow Breakpoint Test: 21


F-statistic 1.189433 Probability 0.328351
Log likelihood ratio 3.992003 Probability 0.262329

(f) Explain the results and implications of the following Ramsey RESET test. (Note that the
dummy variables have been omitted from the original model).

Ramsey RESET Test:


F-statistic 0.000200 Probability 0.988806
Log likelihood ratio 0.000219 Probability 0.988189

Test Equation:
Dependent Variable: INV
Method: Least Squares
Date: 05/15/02 Time: 13:07
Sample: 1 40
Included observations: 40
Variable Coefficient Std. Error t-Statistic Prob.
C 17.81458 8.199161 2.172732 0.0365
V 0.015226 0.006706 2.270632 0.0293
K 0.144467 0.065596 2.202383 0.0341
FITTED^2 -2.87E-05 0.002028 -0.014128 0.9888
R-squared 0.809773 Mean dependent var 72.59075
Adjusted R-squared 0.793921 S.D. dependent var 47.24981
S.E. of regression 21.44950 Akaike info criterion 9.063919
Sum squared resid 16562.91 Schwarz criterion 9.232807
Log likelihood -177.2784 F-statistic 51.08255
Durbin-Watson stat 1.106556 Prob(F-statistic) 0.000000

Note: We can have similar questions using results from eviews to check for autocorrelation
and heteroscedasticity (Breusch Godfrey test and White test).
PROBLEM 3:
You are using an econometric model to study the dependence of the annual salaries of CEOs
(Chief Executive Officers) of major private companies on some variables. The sample data
consist of observations for 60 private firms which include the following variables:
SALi : the annual salary of the CEO of firm i, measured in thousands of dollars;
ARi : the annual total sales revenues of firm i, measured in millions of dollars;
MVi: the market value of firm i, measured in millions of dollars.
EMi: the number of years the CEO has been employed with firm i;
AGEi: the age of the CEO of firm i, in years.
The regression model you propose is:
ln SALi   1   2 ln ARi   3 ln MVi   4 EM i   5 EM i2   6 AGEi   7 AGEi2  u i
In which: (lnXi) denotes the natural logarithm of Xi. EM2 and AGE2 are the squares of
corresponding variables, ui is stochastic disturbance.
Using the data, you estimate the following regression models (estimated standard errors
in parentheses below the coefficient estimates):
(1) lnSALi (hat)=5.572 + 0.182lnARi + 0.102lnMVi + 0.046EMi - 0.00122EMi2 – 0.042AGEi + 0.00033AGEi2
se (0.0412) (0.0493) (0.0142) (0.000476) (0.0412) (0.00036)

RSS= 42.060; TSS= 64.646


(2) lnSALi (hat)= 4.369 + 0.1646lnARi + 0.1085lnMVi + 0.04512EMi - 0.00121EMi2
RSS= 42.474; TSS= 64.646
1. In the model (1) above, interpret the meaning of each estimated coefficients ˆ 2 , ˆ3
Does each independent variable AR or MV affect the salaries of CEOs?
2. In the model (1), by how much the model can explain for the variation of salaries of
CEOs? Is it correct to say that all independent variables of the model (1) simultaneously do not
explain for the variation of the salaries of CEOs?
3. In the model (1), test the hypothesis that coefficients of AR and MV are equal given that:
cov(ˆ 2 , ˆ3 )  0.001473
4. State the coefficient restrictions that are imposed on regression equation (1) in estimating
model (2) above? Conduct a test of these coefficient restrictions and state the meaning of this
test? Based on the outcome of the test, would you choose equation (2) or equation (1)?
5. What are the implications of introducing the squared terms of EM and AGE in the model
(1)? Present the procedure to use F-test to test the hypothesis that we can drop out two squared
terms EM2 and AGE2 from model (1) (use the form of population regression model).
PROBLEM 4:
You want to study the dependence of beer expenditures of employees in a company on
their incomes, ages and sexes. You have collected a random sample of observations on 40 office
employees, 20 of whom are females and 20 of whom are males. Here is the description of
variables in the data set:
BEi: the annual beer expenditures of employee i, measured in dollars per year.
INCi: the annual income of employee i, in thousands of dollars per year.
AGEi: the age of employee i, in years.
SEXi: the dummy variable, SEXi = 1 if employee i is female and SEXi
= 0 if employee i is male.
You propose the following model (model (1)):
BEi   1   2 INCi   3 AGEi   4 SEX i   5 SEX i * INC i   6 SEX i * AGEi  u i

Using OLS method in EVIEWS, you obtain the following results:


Result (1)
Dependent variable: BE
Included observations: 40

Variable Coefficient Std. Error t-Statistic Prob.

C 489.8631 73.85524 6.632747 0.0000

INC 0.002893 0.000775 3.734180 0.0007

AGE -10.07924 2.229676 -4.520493 0.0001

SEX -265.8574 113.3658 -2.345129 0.0250

SEX*INC -0.001029 0.000971 -1.059491 0.2968

SEX*AGE 4.231494 3.648383 1.159827 0.2542

R-squared 0.6470

Result (2)
BEi = 459.21+ 0.0023 INCi - 8.42 AGEi -169.87 SEXi R2=0.6294
Result (3)
BEi = 342.88+ 0.00238 INCi - 7.575 AGEi R2= 0.3292
1. Write down the sample regression model of model (1) based on the result (1)? Write
down the population regression model and sample regression model for male and female
employees and explain the meaning of the estimated regression coefficients?
Male: BE = β1 + β2*INC + β3*AGE
Female: BE = (β1 + β4) +(β2 + β5)*INC + (β3 + β6)*AGE
- β4: if we fixed income and age, the expenditure for beer of male and female is β4
- The rate of change of expenditure of beer with respect to income for male is β 2, for female
is β2 + β5.  β5 is the difference between rate of change of BE with respect to income of male
and female. It means when income increase $1000 per year, the increase in bees’s
expenditure of male and female deffers by β5 (AGE fixed)
- β6 is the difference between the rate of change of with respect to Age of male and female. It
means when age increase one year, the increase in bees’s expenditure of male and female
differs by β6.
- In particular:
β5 = -0.001029: If AGE is fixed, Income increases $1000/year, the increase in
expenditure for beer is -0.001029 male more than that of female
β6 = 4.231494: If AGE increases by 1 year, the decrease in expenditure for beer is
$4.231494/year of female less than that of male
β2 = 0.002893 is the rate of change of beer with respect to income for male, it
means if Income increases by $1000/year, the BE will increase $0.002893/year.
β3 = -10.07924 is the rate of change of beer with respect to income for male, it
means if Age increases by one year, the BE will decrease $10.07924/year.
β2 + β5 = 0.001864 is the rate of change of beer with respect to income for female,
it means if Income increases by $1000/year, the BE will increase $0.001864/year.
β3 + β6 = -5.847746 is the rate of change of beer with respect to income for
female, it means if Age increase one year, the BE will decrease $5.847746/year.
- Expect the sign of β4: negative because sex = 1 if female and = 0 if male, it means the
difference between male’s expenditure and female’s expenditure is β4 or β4 is the expenditure
of female over male, so β4 should be negative.
2. Using result (1), for male employees, how the expenditures for beer change if their
income increases 1000USD/year? Answer the same question for female employees given that:
cov(ˆ 2 , ˆ5 )  0  XD confidence interval for β2, and β2 + β5
 For male (female) when the income increases 1000USD/year, the expenditure for beer
change from … to …
3. In the model (1), state the null and the alternative hypothesis if you want to test that the
models for the expenditures of beer for male and female are not different in slope
coefficients of both INC and AGE. In other words, you want to conduct the joint test (test cùng
một lúc) of hypothesis of equal slope coefficients of male and female for INC and equal slope
coefficients of male and female for AGE. Perform this test using appropriate information given
above.  test for β5 = β6 = 0 (Dropping test from model (1) to model (2))
4. Using the results above to test the hypothesis that the variable SEX does not affect the
annual expenditures for beer.  hệ số gắn vs biến SEX sẽ = 0 (test for β4 = β5 = β6 = 0)
 droping test from model (1) to model (3)
5. Given that d-DW statistic is 1.92. Using this value to test the problem that can be existed
in the model.
PROBLEM 5:

In order to explain the US defense budget, you are using the data from 1962 to 1981 with
the following variables (all measured in billions USD) and estimate the corresponding model
(Model 1): (Use α=0.05 for references) 

Yt: Defense budget outlay for year t


X2t: GNP for year t
X3t: US military sales in year t
X4t: Aerospace industry sales in year t
D1t: Dummy variable presenting the military conflict involving more than 100,000
troops; D1t = 1 if more than 100,000 troops are involved and equal to 0 if fewer than
100,000 troops are involved.

Dependent Variable: Y                          Sample: 1962 1981


Method: Least Squares                         Included observations:
20
Variable Coefficient Std. Error t-Statistic Prob.  
C 21.40251 1.496947 14.29744 0.0000
D1 -48.21987 6.871544 -7.017328 0.0000
X2 0.013879 0.003207 4.328062 0.0008
X3 0.073146 0.203805 0.358902 0.7254
X4 1.389753 0.130197 10.67423 0.0000
X4*D1 1.540792 0.325005 4.740818 0.0004
X2*D1 0.022406 0.005781 3.876038 0.0019
R-squared 0.996366     Mean dependent var 83.86000
Adjusted R-squared 0.994688     S.D. dependent var 28.97771
S.E. of regression 2.111972     Akaike info criterion 4.602338
Sum squared resid 57.98554     Schwarz criterion 4.950845
Log likelihood -39.02338     F-statistic 593.9815
Durbin-Watson stat 2.233771     Prob(F-statistic) 0.000000

1. [10] Explain the meaning of each estimated coefficient and R in the above model.
2

2. [10] Test for significance of each independent variable and test for overall significance of
the model.
3. [15] Conduct the test of autocorrelation in the model using the information above. State
clearly the conditions to apply this test. If those conditions are not met, name other tests
you can use instead.
4. [10] When GNP increases by 1 bil USD (other variables unchanged), what is the
confidence interval of the difference in the changing levels of defense budget between
the cases of there are more than 100,000 or fewer than 100,000 troops involved in the
military conflict? 
5. [20] For the case when there are fewer than 100,000 troops involving in the conflict (this
condition indicates that we are concerning on the coefficients of X2 and X4 only), if we
simultaneously increase X2 and X4 by 1 billion USD, test the proposition that the
defense budget will increase 1.4 billion USD. What is the confidence interval for the
increase in the level of defense budget in this case? (The covariance between two
estimated coefficients of 2 variables X2 and X4 is -0.00036).
6. [15] Do you think that the military budget does not depend on the number of troops
involving in the conflict given that if you regress Y on X2, X3 and X4 (with intercept),
you get R = 0.971 and RSS = 461.28?
2

7. [20] Given the information below, test for all possible problems in the model 1 above. In
each test specify clearly type of test, type of problem, the statistic used, null and
alternative hypothesis and conclusion about the problem. 

Result (1)
Result (2)

White Heteroskedasticity Test: (No cross term)


F-statistic 2.379399   0.114212
Probability
Obs*R- 15.31799   0.168397
squared Probability

Result (3)
  
Breusch-Godfrey Serial Correlation LM Test: AR(2)
F-statistic 1.950537 Probability 0.188349
Obs*R-squared 5.235963 Probability 0.072950

Result (4)

Ramsey RESET Test: 


F-statistic 2.110154 Probabilit 0.119102
y
Log likelihood 7.432899 Probabilit 0.059308
ratio y

You might also like