Test Exer 6

MOOC Econometrics
Test Exercise 6
In answering the questions below, you should use the seasonally adjusted production data denoted by ‘toyota-sa’
and ‘other-sa’. We will denote these variables by y = toyota-sa and x = other-sa.
(a) Make time series plots of the variables yt and xt, and also of the share of Toyota in all produced passenger cars,
that is yt/(yt + xt). What conclusions do you draw from these plots?
Answer:
Chart Title
700000.0
600000.0
500000.0
400000.0
300000.0
200000.0
100000.0
0.0
22
50
239
1
8
15
29
36
43
57
64
71
78
85
92
99
106
113
120
127
134
141
148
155
162
169
176
183
190
197
204
211
218
225
232
246
y TOYOTA_SA x OTHER_SA
the data shows that there is an additive seasonal trend for the production of the other cars as between 1991July
and 1991 February there was an increase in the production of the cars.
The data shows that there is a multiplicative trend in the production of the Toyota cars and between 1990 August
and 1991 February they also produced more cars
y/(y+x)
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
17
201
1
9
25
33
41
49
57
65
73
81
89
97
105
113
121
129
137
145
153
161
169
177
185
193
209
217
225
233
241
249
Market Share
(b) (i) Perform the Augmented Dickey-Fuller (ADF) test for yt. In the ADF test equation, include a constant (α) and
three lags of ∆yt, as well as the variable of interest, yt−1. Report the coefficient of yt−1 and its standard error and
t-value, and draw your conclusion.
ANSWER:
∆yt = α +yt + δyt−1 + δyt−2+ei.
H0 : δ=0, not stationery
H1 : δ≠0, stationery
Toyota: yt-1 = -0.083219

SE = 0.0368
t- stat = -2.262
now t- stat for toyota are greater than the critical value of -2.9 hence we fail to reject the null hypothesis and
conclude it is not stationery.
(ii) Perform a similar ADF test for xt.
ANSWER:
H0 : δ=0, not stationery

H1 : δ≠0, stationery
Other: xt-1 = -0.06962

SE = 0.03067
t-stat = -2.106
now t- stat for other are greater than the critical value of -2.9 hence we fail to reject the null hypothesis and
conclude it is not stationery.
(c) Perform the two-step Engle-Granger test for cointegration of the time series yt and xt. In step 1, regress yt on a
constant and xt. In step 2, perform a regression of the residuals et in the model ∆et = α + ρet−1 + β1∆et−1 + β2∆et−2
+ β3∆et−3 + ωt. What is your conclusion?
ANSWER:
H0: not cointergrated
H1: cointergrated
Yt = 26786.43 +0.45xt +et
Yt -26786.43- 0.45xt = et
regressing Δet
Δet = 29.99 -0.29et-1 -0.4Δet-2-0.104Δet-3 +wt
Now the standard error is 0.068

T statistic is -4.306
The critical value from the table of 5.1 in Engle-Granger is -3.4
Now -4.3 <-3.4 , we reject the null hypothesis and conclude the data is cointergrated.
(d) Construct the first twelve sample autocorrelations and sample partial autocorrelations of ∆yt and use the
outcomes to motivate an AR(12) model for ∆yt. Check that only the lagged terms at lags 1 to 5, 10, and
12 are significant, and estimate the following model: ∆yt t
(recall that the estimation sample is Jan 1980 - Dec 1999).
ANSWER:
∆yt ( AR(12)):
∆yt = 619.04 -0.62∆yt-1 -0.30∆yt-2 -0.26∆yt-3 -0.27∆yt-4 – 0.23∆yt-5 - 0.12 ∆yt-6 - 0.13∆yt-7 +0.04∆yt-8 +0.04∆yt-9 –
0.26∆yt-10 - 0.04∆yt-11 + 0.22∆yt-12
From checking the significance of the values with our p-value the lagged values which are significant should be
lesser than the p value that is the lagged values from 1 to 5 , 10 and 12 hence they are significant.
(e) Extend the model of part (d) by adding the Error Correcion (EC) term (yt −0.45xt), that is, estimate the ECM
∆yt (estimation sample is Jan 1980
- Dec 1999). Check that the EC term is significant at the 5% level, but not at the 1% level.
ANSWER:
∆yt = 4728.01 -0.15( yt-1 -0.45 Xt-1) -0.52∆yt-1 -0.19∆yt-2 -0.16∆yt-3 – 0.18∆yt-4 - 0.13 ∆yt-5 - 0.27∆yt-10 +0.25∆yt-12
now
-0.15( yt-1 -0.45 Xt-1):
T value = -2.16
P value = 0.03
SE =0.07
f) RMSE:- AR = 16992, ECM=18205 HAE:- AR=14703, ECM = 1556

Test Exer 6

Uploaded by

Copyright:

Available Formats

Test Exer 6

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Test Exer 6

Uploaded by

Copyright:

Available Formats

What conclusions can be drawn from the time series plots of yt, xt, and their share?

What are the results of the Augmented Dickey-Fuller tests for yt and xt?

MOOC Econometrics

Toyota: yt-1 = -0.083219

(ii) Perform a similar ADF test for xt.

H0 : δ=0, not stationery

Other: xt-1 = -0.06962

H0: not cointergrated

Yt = 26786.43 +0.45xt +et

Δet = 29.99 -0.29et-1 -0.4Δet-2-0.104Δet-3 +wt

Now the standard error is 0.068

The critical value from the table of 5.1 in Engle-Granger is -3.4

0.26∆yt-10 - 0.04∆yt-11 + 0.22∆yt-12

-0.15( yt-1 -0.45 Xt-1):

f) RMSE:- AR = 16992, ECM=18205 HAE:- AR=14703, ECM = 1556

You might also like