Table 14 (Cont.) - Random effects ordered Probit
|
__________Moody’s__________ |
__________S&P__________ |
________________Fitch________________ | ||||
|
(1) |
(2) |
(3) |
(4) |
(5) |
(6) | |
|
Constant |
8.13 |
7.00 |
3.22 |
7.63 |
2.46 |
3.71 |
|
Cut1 |
1.00 |
1.00 |
1.00 |
1.00 |
1.00 |
1.00 |
|
Cut2 |
2.00 |
2.06 |
2.19 |
2.16 |
2.35 |
2.38 |
|
Cut3 |
3.40 |
3.36 |
4.12 |
4.07 |
3.33 |
3.43 |
|
Cut4 |
4.94 |
5.01 |
5.34 |
5.34 |
4.64 |
4.82 |
|
Cut5 |
5.94 |
6.14 |
7.11 |
7.19 |
5.77 |
5.93 |
|
Cut6 |
7.09 |
7.35 |
9.15 |
9.32 |
7.51 |
7.54 |
|
Cut7 |
8.65 |
8.92 |
10.75 |
10.80 |
9.13 |
9.02 |
|
Cut8 |
10.72 |
10.75 |
13.11 |
12.92 |
10.80 |
10.81 |
|
Cut9 |
11.76 |
11.82 |
14.59 |
14.30 |
11.82 |
12.02 |
|
Cut10 |
12.97 |
13.13 |
15.46 |
14.99 |
12.92 |
13.10 |
|
Cut11 |
14.25 |
14.49 |
17.49 |
16.59 |
15.30 |
15.42 |
|
Cut12 |
15.50 |
15.72 |
18.96 |
18.00 |
16.99 |
17.52 |
|
Cut13 |
17.62 |
17.50 |
21.51 |
19.99 |
17.63 |
18.42 |
|
Cut14 |
19.11 |
18.86 |
22.72 |
21.07 |
19.85 |
20.87 |
|
Cut15 |
20.60 |
20.26 |
24.54 |
23.00 |
22.11 |
23.07 |
|
Cut16______________ |
_______21.64 |
21.26 |
27.07 |
25.69 |
24.06 |
25.04 |
|
LogLik |
-566.33 |
-578.24 |
-514.45 |
-531.22 |
-537.09 |
-533.09 |
|
Observations |
________551 |
557 |
564 |
565 |
553 |
564 |
|
Equal differences $ |
29.26 (0.009) |
19.91 (0.133) |
52.21 (0.000) |
59.68 (0.000) |
68.57 (0.000) |
70.23 (0.000) |
|
Jump& |
[7-8] |
[7-8] |
[9-10] |
[12-13] | ||
|
Different Slopes# |
[2-3 ,5-6, 7-8, |
[2-3, 5-6, 7-8, |
[10-11, 13-14, |
[10-11, 11- 12, | ||
Test*18.22 (0.149) 12.22 (0.510) 19.23 (0.116) 14.02 (0.300) 22.03 (0.037) 16.69 (0.214)
Notes: The t statistics are in parentheses. *, **, *** - statistically significant at the 10, 5, and 1 per cent $The null is that the differences
between categories is equal for all categories. The test statistic is to be compared to a Qui-Square with 14 degrees of freedom. & Identifies
two cut points that have a irregular difference. # Identifies a cluster of categories that seem to have a higher slope (increase difficulty in
transition between adjacent notches). *The null is that, excluding the jump point, within the two identified clusters the slopes are equal. The
test statistic is to be compared to a Qui-Square with either 13 degrees of freedom (if only a jump or different slopes was identified) or 12
degrees of freedom (if both where identified). The p-value is in brackets.
For the correspondence between the ratings and the cut-off points see note to Table 13.
48