1.
Create a main folder on desktop and rename it as your ‘Examination Roll
number’ (Complete 11 Digits Number).
2. Both the data Excel file and data file generated in Gretl should also be saved,
with the name ‘Data file1’ and ‘Data file2’ respectively. All files should be
saved in the main folder.
3. Both the question paper and the evaluation sheet are to be submitted to the
teacher-on-duty in the computer lab.
4. The Gretl generated outputs are to be copied and pasted question-wise in one
word file, which is to be named as ‘Answer file’. The answers are to be
pasted/written in the ‘Answer File’ in the same order in which questions are
mentioned below.
Answer a)
Model 1: OLS, using observations 1-32
Dependent variable: AuctionPrice
Coefficient Std. Error t-ratio p-value
const −1336.05 175.272 −7.623 <0.0001 ***
AgeofClock 12.7414 0.912356 13.97 <0.0001 ***
Noofbidders 85.7641 8.80199 9.744 <0.0001 ***
Mean dependent var 1328.094 S.D. dependent var 393.6495
Sum squared resid 525462.2 S.E. of regression 134.6083
R-squared 0.890614 Adjusted R-squared 0.883070
F(2, 29) 118.0585 P-value(F) 1.16e-14
Log-likelihood −200.7068 Akaike criterion 407.4136
Schwarz criterion 411.8108 Hannan-Quinn 408.8711
Est Yi = −1336.05 +12.7414 Age of Clock +85.7641No. of bidder
Interpretation and significance of the slope coefficient of age of clock:
Test of Slope of age of Clock: Sinc the p-value is less than 5%, therefore we reject
the NH that coefficient of age of clock is zero. The variable age has significant
impact on the auction price of clock.
Answer b)
Model 2: OLS, using observations 1-32
Dependent variable: AuctionPrice
Coefficient Std. Error t-ratio p-value
const −8660.53 657.727 −13.17 <0.0001 ***
l_AgeofClock 1849.69 127.491 14.51 <0.0001 ***
Noofbidders 85.6242 8.50809 10.06 <0.0001 ***
Mean dependent var 1328.094 S.D. dependent var 393.6495
Sum squared resid 491538.9 S.E. of regression 130.1907
R-squared 0.897676 Adjusted R-squared 0.890619
F(2, 29) 127.2069 P-value(F) 4.41e-15
Log-likelihood −199.6390 Akaike criterion 405.2780
Schwarz criterion 409.6752 Hannan-Quinn 406.7356
Answer c)
Model 3: OLS, using observations 1-32
Dependent variable: l_AuctionPrice
Coefficient Std. Error t-ratio p-value
const −1.31977 0.388293 −3.399 0.0020 ***
l_AgeofClock 1.41845 0.0711387 19.94 <0.0001 ***
l_Noofbidders 0.648501 0.0447974 14.48 <0.0001 ***
Mean dependent var 7.148258 S.D. dependent var 0.301306
Sum squared resid 0.154445 S.E. of regression 0.072977
R-squared 0.945122 Adjusted R-squared 0.941338
F(2, 29) 249.7241 P-value(F) 5.26e-19
Log-likelihood 39.93246 Akaike criterion −73.86492
Schwarz criterion −69.46771 Hannan-Quinn −72.40737
Answer d)
Answer
White's test for heteroskedasticity
OLS, using observations 1-32
Dependent variable: uhat^2
coefficient std. error t-ratio p-value
---------------------------------------------------------------
const 30945.6 134716 0.2297 0.8201
AgeofClock −590.720 1575.70 −0.3749 0.7108
Noofbidders 4758.94 10956.3 0.4344 0.6676
sq_AgeofClock 3.57774 4.83940 0.7393 0.4664
X2_X3 −45.7695 40.7857 −1.122 0.2720
sq_Noofbidders 104.686 377.769 0.2771 0.7839
Unadjusted R-squared = 0.123966
Test statistic: TR^2 = 3.966897,
with p-value = P(Chi-square(5) > 3.966897) = 0.554192
NH: There is no heteroscedasticity. AH: the errors are heteroscedastic.
Test statistic n*R2 And it follows chi-squared distribution.
Errors are homoscedastic.
Answer
Normality of Residuals
Jarque Bera test for normality of residuals:
NH:…… AH:…..
Test for null hypothesis of normal distribution:
Chi-square(2) = 1.227 with p-value 0.54134
Answer:
Graph of Actual and fitted value
99% CI of slope
t(29, 0.005) = 2.756
coefficient 99 confidence interval
const -1336.05 [-1819.17, -852.931]
AgeofClock 12.7414 [10.2266, 15.2562]
Noofbidders 85.7641 [61.5024, 110.026]