APPENDIX C
Table C.1: Johansen’s Cointegration Rank Test Statistics for AI system applied to subsamples of
Italian survey data
|
Gender |
Education |
Income | |||||||||||
|
Male |
Female |
High |
Low |
Above pov. |
thresh. |
Below |
pov. thresh. |
95% critical values | |||||
|
H0 |
LR |
LR |
LR |
LR |
LR |
LR |
λ trace | ||||||
|
r=0 |
157.80 |
[0.000] |
136.02 |
[0.000] |
162.69 |
[0.000] |
132.75 |
[0.000] |
136.50 |
[0.000] |
126.97 |
[0.000] |
103.68 |
|
r=1 |
90.89 |
[0.002] |
75.98 |
[0.058] |
106.86 |
[0.000] |
77.64 |
[0.043] |
76.37 |
[0.050] |
76.20 |
[0.058] |
76.81 |
|
r=2 |
39.74 |
[0.488] |
41.84 |
[0.386] |
56.88 |
[0.026] |
50.76 |
[0.095] |
49.22 |
[0.126] |
46.74 |
[0.193] |
53.94 |
|
r=3 |
19.40 |
[0.761] |
21.48 |
[0.633] |
27.62 |
[0.262] |
25.72 |
[0.362] |
27.42 |
[0.279] |
21.91 |
[0.604] |
35.07 |
|
r=4 |
9.95 |
[0.649] |
9.77 |
[0.667] |
11.56 |
[0.497] |
9.59 |
[0.683] |
9.08 |
[0.730] |
11.61 |
[0.492] |
20.16 |
|
r=5 |
3.44 |
[0.512] |
3.68 |
[0.473] |
2.27 |
[0.724] |
3.86 |
[0.445] |
3.81 |
[0.453] |
4.21 |
[0.394] |
9.14 |
Age
|
<35 |
35-49 |
50-65 |
>65 |
95% critical values | |
|
H0 |
LR |
LR |
LR |
LR |
λ trace |
|
r=0 |
127.01 [0.000] |
134.93 [0.000] |
167.21 [0.000] |
131.27 [0.000] |
103.68 |
|
r=1 |
80.33 [0.025] |
81.26 [0.021] |
79.98 [0.027] |
79.82 [0.028] |
76.81 |
|
r=2 |
45.46 [0.236] |
49.86 [0.112] |
51.58 [0.081] |
45.43 [0.237] |
53.94 |
|
r=3 |
19.85 [0.735] |
27.57 [0.264] |
30.17 [0.158] |
23.84 [0.477] |
35.07 |
|
r=4 |
8.21 [0.802] |
12.42 [0.420] |
10.16 [0.629] |
9.94 [0.650] |
20.16 |
|
r=5 |
2.62 [0.659] |
5.05 [0.288] |
3.51 [0.500] |
3.86 [0.444] |
9.14 |
Notes: r is number of cointegrating relations. We report Johansen trace tests for evaluating the identification condition
of AI model. These values are estimated by optimal length lags with restricted intercepts and no trends in six endogenous
variables w1 , w2 , logp1 , logp2 , logp3 and log(Y /p). Seasonal centred dummies are included. P-values of tests are shown
in square brackets.
33