Nafiah Rafiqah Rahayu

3SE4 / 14.8281
Latihan Soal Gujarati
21.19 Consider the dividends and profits time series given in Table 21.1.
Since dividends depend on profits, consider the following simple model:
Dividendst = 1 + 2Profits + ut
a. Would you expect this regression to suffer from the spurious regression
phenomenon? Why?
1. Uji Root Test pada Group Deviden dan Profit di level
Null Hypothesis: Unit root (individual unit root process)
Series: DIVIDEND, PROFITS
Date: 12/20/16 Time: 15:03
Sample: 1 88
Exogenous variables: None
Automatic selection of maximum lags
Automatic lag length selection based on SIC: 1 to 2
Total number of observations: 171
Cross-sections included: 2
Method
ADF - Fisher Chi-square
ADF - Choi Z-stat

Statistic
0.32842
3.39145

Prob.**
0.9879
0.9997

** Probabilities for Fisher tests are computed using an asymptotic Chi

-square distribution. All other tests assume asymptotic normality.
Intermediate ADF test results UNTITLED

Series
DIVIDEND
PROFITS

Prob.
0.9999
0.8486

Lag
2
1

Max Lag
11
11

Obs
85
86

Hasilnya Gagal tolak Ho berarti data profits dan deviden secara

terintegrasi belum stasioner di level

2. Lalu dilihat pada 1st Difference apa sudah stasioner

1

Null Hypothesis: Unit root (individual unit root process)

Series: DIVIDEND, PROFITS
Date: 12/20/16 Time: 15:05
Sample: 1 88
Exogenous variables: None
Automatic selection of maximum lags
Automatic lag length selection based on SIC: 0 to 1
Total number of observations: 171
Cross-sections included: 2
Method
ADF - Fisher Chi-square
ADF - Choi Z-stat

Statistic
74.9971
-7.76279

Prob.**
0.0000
0.0000

** Probabilities for Fisher tests are computed using an asymptotic Chi

-square distribution. All other tests assume asymptotic normality.
Intermediate ADF test results D(UNTITLED)

Series
D(DIVIDEND)
D(PROFITS)

Prob.
0.0000
0.0000

Lag
1
0

Max Lag
11
11

Obs
85
86

Hasilnya Tolak Ho yang berarti data profit dan deviden terintegrasi

stasioner di 1st difference
3. Lalu regresikan Profit (x) dan Deviden (y) di level untuk menjadi
persamaan Long Run
Dependent Variable: DIVIDEND
Method: Least Squares
Date: 12/20/16 Time: 15:08
Sample: 1 88
Included observations: 88
Variable

Coefficient

Std. Error

t-Statistic

Prob.

C
PROFITS

-13.31143
0.628156

7.362611
0.052674

-1.807977
11.92531

0.0741
0.0000

R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
F-statistic
Prob(F-statistic)

0.623159
0.618777
23.68747
48254.27
-402.3703
142.2130
0.000000

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.
Durbin-Watson stat

69.16477
38.36447
9.190234
9.246538
9.212918
0.071211

Nilai Durbin Watson < R2 maka terjadi spurious regression (regresi

tidak bermakna)

b. Are Dividends and Profits time series cointegrated? How do you test for
this explicitly? If, after testing, you find that they are cointegrated, would
your answer in a change?

Null Hypothesis: ECT has a unit root

Exogenous: Constant
Lag Length: 1 (Automatic - based on SIC, maxlag=11)

Augmented Dickey-Fuller test statistic

Test critical values:
1% level
5% level
10% level

t-Statistic

Prob.*

-1.547343
-3.508326
-2.895512
-2.584952

0.5050

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(ECT)
Method: Least Squares
Date: 12/20/16 Time: 15:29
Sample (adjusted): 3 88
Included observations: 86 after adjustments
Variable

Coefficient

Std. Error

t-Statistic

Prob.

ECT(-1)
D(ECT(-1))
C

-0.043191
0.365150
0.149913

0.027913
0.103429
0.644144

-1.547343
3.530427
0.232733

0.1256
0.0007
0.8165

R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
F-statistic
Prob(F-statistic)

0.138697
0.117943
5.965967
2954.199
-274.1040
6.682823
0.002037

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.
Durbin-Watson stat

0.266884
6.352322
6.444280
6.529897
6.478737
2.008585

Bagaimana menguji kointegrasi?

1. Data non stasioner pada level -> sudah terpenuhi
2. Data secara bersama-sama terintegrasi di 1st difference ->
sudah terpenuhi
3. Diuji apakah error sudah normal di level
Maka save residualnya dan uji apa sudah stasioner di level
Ternyata Hasil Uji ADF gagal tolak Ho sehingga data tidak stasioner
di level
Sehingga Profit dan Deviden tidak terkointegrasi

c. Employ the error correction mechanism (ECM) to study the short and
long-run behavior of dividends in relation to profits.
Persaman Long Run
Dependent Variable: DIVIDEND
Method: Least Squares
Date: 12/20/16 Time: 15:08
Sample: 1 88
Included observations: 88
Variable

Coefficient

Std. Error

t-Statistic

Prob.

C
PROFITS

-13.31143
0.628156

7.362611
0.052674

-1.807977
11.92531

0.0741
0.0000

R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
F-statistic
Prob(F-statistic)

0.623159
0.618777
23.68747
48254.27
-402.3703
142.2130
0.000000

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.
Durbin-Watson stat

69.16477
38.36447
9.190234
9.246538
9.212918
0.071211

Persamaan Short Run

Dependent Variable: D(DIVIDEND)
Method: Least Squares
Date: 12/20/16 Time: 15:33
Sample (adjusted): 2 88
Included observations: 87 after adjustments
Variable

Coefficient

Std. Error

t-Statistic

Prob.

C
D(PROFITS)
ECT(-1)

1.348006
-0.023179
-0.003050

0.196477
0.021000
0.008329

6.860876
-1.103799
-0.366166

0.0000
0.2728
0.7152

R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
F-statistic
Prob(F-statistic)

0.016812
-0.006597
1.802361
272.8746
-173.1727
0.718194
0.490602

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.
Durbin-Watson stat

1.310345
1.796446
4.049947
4.134978
4.084186
0.597216

Error tidak signifikan pada model ECM sehingga model tidak cocok

16

Series: Residuals
Sample 2 88
Observations 87

14
12
10
8
6
4
2

Mean
Median
Maximum
Minimum
Std. Dev.
Skewness
Kurtosis

2.22e-16
-0.191087
5.853065
-4.355732
1.781280
0.426892
3.639756

Jarque-Bera
Probability

4.126102
0.127066

0
-4

-3

-2

-1

Error berdistribusi normal pada short run

d. If you examine the Dividends and Profits series individually, do they
exhibit stochastic or deterministic trends? What tests do you use?
Profit
Null Hypothesis: PROFITS has a unit root
Exogenous: None
Lag Length: 1 (Automatic - based on SIC, maxlag=11)

Augmented Dickey-Fuller test statistic

Test critical values:
1% level
5% level
10% level

t-Statistic

Prob.*

0.621075
-2.592129
-1.944619
-1.614288

0.8486

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(PROFITS)
Method: Least Squares
Date: 12/21/16 Time: 00:16
Sample (adjusted): 3 88
Included observations: 86 after adjustments
Variable

Coefficient

Std. Error

t-Statistic

Prob.

PROFITS(-1)
D(PROFITS(-1))

0.004444
0.279516

0.007155
0.105926

0.621075
2.638779

0.5362
0.0099

R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
Durbin-Watson stat

0.058818
0.047614
9.119260
6985.516
-311.1103
1.999075

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.

1.696512
9.344435
7.281636
7.338713
7.304607

Null Hypothesis: D(PROFITS) has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=11)
t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

Test critical values:
1% level
5% level
10% level

-6.827867
-2.592129
-1.944619
-1.614288

0.0000

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(PROFITS,2)
Method: Least Squares
Date: 12/21/16 Time: 00:18
Sample (adjusted): 3 88
Included observations: 86 after adjustments
Variable

Coefficient

Std. Error

t-Statistic

Prob.

D(PROFITS(-1))

-0.708427

0.103755

-6.827867

0.0000

R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
Durbin-Watson stat

0.354200
0.354200
9.086250
7017.594
-311.3073
2.005202

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.

0.011628
11.30669
7.262961
7.291500
7.274447

Data tidak stasioner di level namun stasioner di 1st difference

sehingga persamaan profit stokastik
Deviden
Null Hypothesis: DIVIDEND has a unit root
Exogenous: None
Lag Length: 2 (Automatic - based on SIC, maxlag=11)

Augmented Dickey-Fuller test statistic

Test critical values:
1% level
5% level
10% level

t-Statistic

Prob.*

3.671337
-2.592452
-1.944666
-1.614261

0.9999

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(DIVIDEND)
Method: Least Squares
Date: 12/21/16 Time: 00:19
Sample (adjusted): 4 88
Included observations: 85 after adjustments
Variable

Coefficient

Std. Error

t-Statistic

Prob.

DIVIDEND(-1)

0.007636

0.002080

3.671337

0.0004

D(DIVIDEND(-1))
D(DIVIDEND(-2))
R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
Durbin-Watson stat

1.060148
-0.516927
0.615326
0.605943
1.125584
103.8890
-129.1383
2.083047

0.096306
0.097561

11.00813
-5.298493

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.

0.0000
0.0000
1.355294
1.793076
3.109137
3.195349
3.143814

Null Hypothesis: D(DIVIDEND) has a unit root

Exogenous: None
Lag Length: 1 (Automatic - based on SIC, maxlag=11)

Augmented Dickey-Fuller test statistic

Test critical values:
1% level
5% level
10% level

t-Statistic

Prob.*

-4.358380
-2.592452
-1.944666
-1.614261

0.0000

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(DIVIDEND,2)
Method: Least Squares
Date: 12/21/16 Time: 00:20
Sample (adjusted): 4 88
Included observations: 85 after adjustments
Variable

Coefficient

Std. Error

t-Statistic

Prob.

D(DIVIDEND(-1))
D(DIVIDEND(-1),2)

-0.267964
0.415006

0.061482
0.100313

-4.358380
4.137122

0.0000
0.0001

R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
Durbin-Watson stat

0.249444
0.240401
1.207236
120.9657
-135.6062
1.934415

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.

0.011765
1.385160
3.237792
3.295266
3.260910

Data tidak stasioner di level namun stasioner di 1st difference

sehingga persamaan deviden stokastik

*e. Assume Dividends and Profits are cointegrated. Then, instead of

regressing dividends on profits, you regress profits on dividends. Is such a
regression valid?
(diasumsikan deviden dan profit terkointegrasi)
1. Regresikan deviden (x) dan profit(y) -> Long Run
Simpan errornya untuk persamaan Short Run
Dependent Variable: PROFITS
Method: Least Squares
Date: 12/20/16 Time: 15:50
Sample: 1 88
Included observations: 88
Variable

Coefficient

Std. Error

t-Statistic

Prob.

C
DIVIDEND

62.68432
0.992045

6.570745
0.083188

9.539911
11.92531

0.0000
0.0000

R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
F-statistic
Prob(F-statistic)

0.623159
0.618777
29.76805
76207.77
-422.4774
142.2130
0.000000

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.
Durbin-Watson stat

131.2989
48.21265
9.647214
9.703517
9.669897
0.105802

DW<R2 maka spurious regression

Persamaan Short Run

Dependent Variable: D(PROFITS)
Method: Least Squares
Date: 12/20/16 Time: 15:52
Sample (adjusted): 2 88
Included observations: 87 after adjustments
Variable

Coefficient

Std. Error

t-Statistic

Prob.

C
D(DIVIDEND)
ECT_PROF(-1)

2.296212
-0.471426
-0.043733

1.240429
0.570168
0.034430

1.851143
-0.826820
-1.270186

0.0677
0.4107
0.2075

R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
F-statistic
Prob(F-statistic)

0.033801
0.010796
9.242117
7175.005
-315.3893
1.469290
0.235941

Mean dependent var

S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.
Durbin-Watson stat

1.673563
9.292413
7.319293
7.404325
7.353533
1.447190

Ketika koefisien regresi Error bertanda negative dan signifikan

maka Persamaan Short Run valid namun pada persamaan ini

Error tidak signifikan sehingga persamaan regresi Short Run tidak

valid (tidak cocok dengan ECM)
21.23 From the U.K. private sector housing starts (X) for the period 1948
to
1984, Terence Mills obtained the following regression results*:

Note: The 5 percent critical value is 2.95 and the 10 percent

critical
value is 2.60.
a. On the basis of these results, is the housing starts time series
stationary or nonstationary? Alternatively, is there a unit root in this time
series? How do you know?
Pengujian Dickey-Fuller
statistik dengan rumus

dilakukan

dengan

menghitung

nilai

Nilai -statistik yang diperoleh kemudian dibandingkan dengan McKinnon Critical Values. Jika -statistik kurang dari -tabel, maka H 0
diterima atau tidak cukup bukti untuk menolak hipotesis bahwa dalam
persamaan mengandung akar unit, artinya data tidak stasioner.
Keputusan |-2.35| < |-2.95| sehingga Gagal tolak H0
Kesimpulan data housing start tidak stasioner sehingga dalam
persamaan mengandung akar unit
b. If you were to use the usual t test, is the observed t value statistically
significant? On this basis, would you have concluded that this time series
is stationary?
Ketika data tidak stasioner akibat adanya akar unit maka signifikansi dari
statistic t diragukan sehingga data tidak stasioner
c. Now consider the following regression results:

where _2 is the second difference operator, that is, the first difference
of the first difference. The estimated value is now statistically
significant.
What can you say now about the stationarity of the time series
in question?
Note: The purpose of the preceding regression is to find out if there
is a second unit root in the time series.
Keputusan |-5.89| > |-2.95| sehingga tolak H0
Kesimpulan data second difference housing starts stasioner sehingga
dalam persamaan tidak mengandung second unit root
21.26 From the data for the period 1971I to 1988IV for Canada, the
following
regression results were obtained:

where M1 = M1 money supply, GDP = gross domestic product, both

measured in billions of Canadian dollars, ln is natural log, and ut
represent
the estimated residuals from regression 1.
a. Interpret regressions 1 and 2.
Regresi 1
DW < R2 sehingga mengalami Spurious Regression , koefisien regresi tidak
dapat diinterpretasikan. Ada indikasi korelasi di errornya
Statistic t cukup tinggi sehingga GDP berpengaruh signifikan terhadap M1
money supply
Regresi 2
DW > R2 sehingga tidak mengalami Spurious Regression , koefisien regresi
dapat diinterpretasikan. Setiap kenaikan perubahan GDP satu persen akan
menambah kenaikan perubahan M1 Money Supply sebesar 0.583 persen
(karena Ln)
10

Statistic t rendah sehingga GDP tidak berpengaruh signifikan terhadap M1

money supply
b. Do you suspect that regression 1 is spurious? Why?
Ya , karena DW < R2 sehingga mengalami Spurious Regression , koefisien
regresi tidak dapat diinterpretasikan. Ada indikasi korelasi di errornya
c. Is regression 2 spurious? How do you know?
DW > R2 sehingga tidak mengalami Spurious Regression , koefisien regresi
dapat diinterpretasikan
d. From the results of regression 3, would you change your conclusion
in b? And why?
Regresi 3 menunjukkan data 1st difference dari error dari regresi 1
Statistic t tidak signifikan sehingga error tidak stasioner pada level
GDP dan Money Supply tidak terkointegrasi
Kesimpulan b tidak diganti
e. Now consider the following regression:

What does this regression tell you? Does this help you decide if regression
1 is spurious or not?
Pada short run datanya tidak spurious regression sehingga dapat
dimaknai koefisien regresinya.Setiap penambahan perubahan GDP satu
persen akan menambah perubahan M1 money suppy 0.734 persen
Namun ternyata statistic t dari error pada levelnya tidak signifikan
sehingga sulit ditentukan apakah regresi 1 spurious atau tidak

11