Collinearity Diagnostics
SQRT R-
Variable VIF VIF Tolerance Squared
---------------------------------------------------- lnsize 4.87 2.21 0.2054 0.7946
nim 1.37 1.17 0.7314 0.2686
lta 1.19 1.09 0.8435 0.1565
eta 1.78 1.33 0.5618 0.4382
cir 1.05 1.03 0.9481 0.0519
gdp 1.28 1.13 0.7793 0.2207
lnnon 4.51 2.12 0.2216 0.7784
---------------------------------------------------- Mean VIF 2.29
Cond
Eigenval Index
---------------------------------
6.0911 | 1.0000 | |
2 | 0.9954 | 2.4737 |
3 | 0.4828 | 3.5520 |
4 | 0.2832 | 4.6376 |
5 | 0.1024 | 7.7120 |
6 | 0.0374 | 12.7689 |
7 | 0.0066 | 30.4174 |
8 | 0.0011 | 73.0061 |
Có thể bạn quan tâm!
-
Kết Quả Ước Lượng Ảnh Hưởng Của Thu Nhập Phi Truyền Thống Đến Rủi Ro Của Ngân Hàng. -
Elsas, R. Hackethal Et Al (2010), The Anatomy Of Bank Diversification, Journal Of Banking And Finance, 3496), Pp, 1274-1287. -
Nguyễn Văn Tiến (2010), “Quản Trị Rủi Ro Trong Kinh Doanh Ngân Hàng”. Nhà Xuất Bản Thống Kê. -
Ảnh hưởng của thu nhập phi truyền thống đến khả năng sinh lời và rủi ro của các ngân hàng ở Việt Nam trong giai đoạn 2005-2013 - 14
Xem toàn bộ 118 trang tài liệu này.
---------------------------------
Condition Number 73.0061
Eigenvalues & Cond Index computed from scaled raw sscp (w/ intercept) Det(correlation matrix) 0.0800
.
Kiểm định phương sai sai số thay đổi
Modified Wald test for groupwise heteroskedasticity in fixed effect regression model
H0: sigma(i)^2 = sigma^2 for al i
12514.43 | |
Prob>chi2 = | 0.5412 |
Kiểm định tự tương quan của phần dư
Wooldridge test for autocorrelation in panel data H0: no first-order autocorrelation
F( 1, 36) = 3.534
Prob > F = 0.0682
Phụ lục 3
Mô hình Pooling OLS
Source | SS | df | MS |
Model | .366234208 | 7 | .052319173 |
Residual | .491038188 | 238 | .002063186 |
Total | .857272396 | 245 | .003499071 |
Number of obs | = | 246 |
F( 7, 238) | = | 25.36 |
Prob > F | = | 0.0000 |
R-squared | = | 0.4272 |
Adj R-squared | = | 0.4104 |
Root MSE | = | .04542 |
Coef. | Std. Err. | t | P>|t| | [95% Conf. | Interval] | |
lnsize | -.0005843 | .0038275 | -0.15 | 0.879 | -.0081245 | .0069559 |
nim | 2.130302 | .26372 | 8.08 | 0.000 | 1.610779 | 2.649826 |
lta | -.6739541 | 1.163781 | -0.58 | 0.563 | -2.96658 | 1.618672 |
eta | -24.35042 | 3.972267 | -6.13 | 0.000 | -32.17571 | -16.52513 |
cir | .3533858 | .3990964 | 0.89 | 0.377 | -.4328267 | 1.139598 |
gdp | 107.7699 | 25.72513 | 4.19 | 0.000 | 57.09186 | 158.4479 |
lnnon | .0123061 | .0026634 | 4.62 | 0.000 | .0070593 | .0175529 |
_cons | -.1248417 | .0566103 | -2.21 | 0.028 | -.2363629 | -.0133205 |
.
Random-effects GLS regression | Number of obs | = | 246 |
Group variable: code | Number of groups | = | 40 |
R-sq: within = 0.1421 | Obs per group: min | = | 2 |
between = 0.6197 | avg | = | 6.2 |
overall = 0.4149 | max | = | 9 |
Wald chi2(7) | = | 91.84 | |
corr(u_i, X) = 0 (assumed) | Prob > chi2 | = | 0.0000 |
Coef. | Std. Err. | z P>|z| | [95% Conf. | Interval] | |||
lnsize | .0017913 | .0041251 | 0.43 0.664 | -.0062938 | .0098763 | ||
nim | 1.641639 | .2726193 | 6.02 0.000 | 1.107315 | 2.175963 | ||
lta | .3402581 | 1.234672 | 0.28 0.783 | -2.079655 | 2.760171 | ||
eta | -20.27894 | 4.024733 | -5.04 0.000 | -28.16727 | -12.3906 | ||
cir | -.3452688 | .4815135 | -0.72 0.473 | -1.289018 | .5984804 | ||
gdp | 102.9646 | 25.43395 | 4.05 0.000 | 53.11498 | 152.8142 | ||
lnnon | .0087438 | .0026622 | 3.28 0.001 | .0035259 | .0139616 | ||
_cons | -.1164649 | .0660718 | -1.76 0.078 | -.2459633 | .0130335 | ||
sigma_u | .02210333 | ||||||
sigma_e | .03920294 | ||||||
rho | .24121182 | (fraction | of | variance due | to | u_i) |
.
Breusch and Pagan Lagrangian multiplier test for random effects
roe[code,t] = Xb + u[code] + e[code,t]
Estimated results:
Var sd = sqrt(Var)
roe .0034991 .0591529
e .0015369 .0392029
u .0004886 .0221033
Test: Var(u) = 0
M.ô hình FEM
chibar2(01) = 19.22
Prob > chibar2 = 0.0000
Number of obs | = | 246 | ||||||
Group variable: code | Number of groups | = | 40 | |||||
R-sq: within | = | 0.1552 | Obs | per | group: | min | = | 2 |
between | = | 0.4916 | avg | = | 6.2 | |||
overall | = | 0.3485 | max | = | 9 | |||
F(7,199) | = | 5.22 | ||||||
corr(u_i, Xb) | = | 0.2664 | Prob > F | = | 0.0000 | |||
Coef. | Std. Err. | t P>|t| | [95% Conf. | Interval] | |||
lnsize | -.0001558 | .0048493 | -0.03 0.974 | -.0097184 | .0094067 | ||
nim | 1.061353 | .3080369 | 3.45 0.001 | .4539172 | 1.668788 | ||
lta | 1.448683 | 1.405181 | 1.03 0.304 | -1.322272 | 4.219639 | ||
eta | -17.26883 | 4.295509 | -4.02 0.000 | -25.73939 | -8.798273 | ||
cir | -1.113798 | .6135187 | -1.82 0.071 | -2.32363 | .0960345 | ||
gdp | 79.31969 | 29.27231 | 2.71 0.007 | 21.59598 | 137.0434 | ||
lnnon | .0059413 | .0028869 | 2.06 0.041 | .0002483 | .0116342 | ||
_cons | -.0246669 | .0867467 | -0.28 0.776 | -.1957277 | .1463939 | ||
sigma_u | .03538221 | ||||||
sigma_e | .03920294 | ||||||
rho | .4489076 | (fraction | of | variance due | to | u_i) |
F test that all u_i=0: F(39, 199) = 3.09 Prob > F = 0.0000
.
Note: the rank of the differenced variance matrix (5) does not equal the number of coefficients being tested (7); be sure this is what you expect, or there may be problems computing the test.
Examine the output of your estimators for anything unexpected and possibly consider scaling your variables so that the coefficients are on a similar scale.
Coefficients
(b) fixed | (B) random | (b-B) Difference | sqrt(diag(V_b-V_B)) S.E. | |
lnsize | -.0001558 | .0017913 | -.0019471 | .0025493 |
nim | 1.061353 | 1.641639 | -.5802859 | .1434066 |
lta | 1.448683 | .3402581 | 1.108425 | .6709076 |
eta | -17.26883 | -20.27894 | 3.010107 | 1.500972 |
cir | -1.113798 | -.3452688 | -.7685288 | .3801972 |
gdp | 79.31969 | 102.9646 | -23.64491 | 14.49076 |
lnnon | .0059413 | .0087438 | -.0028025 | .0011167 |
b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg
Test: Ho: difference in coefficients not systematic
chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 22.14
Prob>chi2 = 0.0005
(V_b-V_B is not positive definite)
.
Tự tương quan của phần dư
Wooldridge test for autocorrelation in panel data H0: no first-order autocorrelation
F( 1, 36) = 11.721
Prob > F = 0.0016
Kiểm định phương sai sai số thay đổi
Modified Wald test for groupwise heteroskedasticity in fixed effect regression model
H0: sigma(i)^2 = sigma^2 for all i
chi2 (40) = 36795.99
Prob>chi2 = 0.0000
.
Phụ lục 4
Source | SS | df | MS |
Model | .125358843 | 7 | .017908406 |
Residual | .131009267 | 238 | .000550459 |
Total | .256368109 | 245 | .0010464 |
Mô hình hồi quy theo biến SDROA Mô hình Pooling OLS
sdroa | Coef. | Std. Err. | t | P>|t| | [95% Conf. | Interval] |
lnsize | -.0083671 | .001977 | -4.23 | 0.000 | -.0122618 | -.0044724 |
nim | 1.445566 | .1362186 | 10.61 | 0.000 | 1.177218 | 1.713914 |
lta | -.2227387 | .6011246 | -0.37 | 0.711 | -1.406943 | .9614657 |
eta | 8.548848 | 2.051785 | 4.17 | 0.000 | 4.506871 | 12.59083 |
cir | .6355336 | .2061442 | 3.08 | 0.002 | .2294333 | 1.041634 |
gdp | 48.58042 | 13.28774 | 3.66 | 0.000 | 22.40382 | 74.75701 |
lnnon | .0098924 | .0013757 | 7.19 | 0.000 | .0071823 | .0126025 |
_cons | .0490729 | .0292408 | 1.68 | 0.095 | -.0085309 | .1066766 |
Mô hình REM
Number of obs = 246
F( 7, 238) = 32.53
Prob > F = 0.0000
= | 0.4890 | |
Adj R-squared | = | 0.4739 |
Root MSE | = | .02346 |
Number of obs | = | 275 | ||||||
Group variable: code | Number of groups | = | 40 | |||||
R-sq: within | = | 0.3600 | Obs | per | group: | min | = | 2 |
between | = | 0.3133 | avg | = | 6.9 | |||
overall | = | 0.3491 | max | = | 9 | |||
Wald chi2(6) | = | 146.09 | ||||||
corr(u_i, X) | = | 0 (assumed) | Prob > chi2 | = | 0.0000 | |||
Coef. | Std. Err. | z P>|z| | [95% Conf. | Interval] | |||
lnsize | .0005293 | .0015783 | 0.34 0.737 | -.0025642 | .0036228 | ||
nim | 1.100125 | .1232902 | 8.92 0.000 | .858481 | 1.34177 | ||
lta | 1.082388 | .6561784 | 1.65 0.099 | -.2036983 | 2.368474 | ||
eta | 5.043126 | 1.70769 | 2.95 0.003 | 1.696115 | 8.390136 | ||
cir | 1.186046 | .2736687 | 4.33 0.000 | .6496655 | 1.722427 | ||
gdp | 55.24661 | 13.79031 | 4.01 0.000 | 28.2181 | 82.27512 | ||
_cons | .0174564 | .0345579 | 0.51 0.613 | -.0502759 | .0851886 | ||
sigma_u | .01395177 | ||||||
sigma_e | .02133582 | ||||||
rho | .2995248 | (fraction | of | variance due | to | u_i) |
Breusch and Pagan Lagrangian multiplier test for random effects
sdroa[code,t] = Xb + u[code] + e[code,t]
Estimated results:
Var sd = sqrt(Var)
sdroa .0010985 .0331438
e .0004552 .0213358
u .0001947 .0139518
Test: Var(u) = 0
Mô hình FEM
chibar2(01) = 34.75
Prob > chibar2 = 0.0000
Fixed-effects (within) regression Number of obs = 275
Group variable: code Number of groups = 40
R-sq: within = 0.3837 Obs per group: min = 2
between = 0.1372 avg = 6.9
overall = 0.2535 max = 9
F(6,229) = 23.76
corr(u_i, Xb) = -0.2368 Prob > F = 0.0000
Coef. | Std. Err. | t P>|t| | [95% Conf. | Interval] | |||
lnsize | -.0012422 | .0019135 | -0.65 0.517 | -.0050124 | .0025281 | ||
nim | 1.00023 | .1301253 | 7.69 0.000 | .7438343 | 1.256626 | ||
lta | 1.145693 | .7055032 | 1.62 0.106 | -.2444145 | 2.5358 | ||
eta | 4.278441 | 1.738839 | 2.46 0.015 | .8522724 | 7.70461 | ||
cir | 2.021628 | .325952 | 6.20 0.000 | 1.37938 | 2.663876 | ||
gdp | 52.01907 | 15.14954 | 3.43 0.001 | 22.16876 | 81.86938 | ||
_cons | .0540468 | .0417874 | 1.29 0.197 | -.0282901 | .1363838 | ||
sigma_u | .02303646 | ||||||
sigma_e | .02133582 | ||||||
rho | .5382705 | (fraction | of | variance due | to | u_i) |
F test that all u_i=0: F(39, 229) = 4.50 Prob > F = 0.0000
.
. hausman fixed random
Coefficients
(b) fixed | (B) random | (b-B) Difference | sqrt(diag(V_b-V_B)) S.E. | |
lnsize | -.0012422 | .0005293 | -.0017715 | .0010818 |
nim | 1.00023 | 1.100125 | -.0998951 | .0416187 |
lta | 1.145693 | 1.082388 | .0633052 | .2591615 |
eta | 4.278441 | 5.043126 | -.7646846 | .3276532 |
cir | 2.021628 | 1.186046 | .8355817 | .1770598 |
gdp | 52.01907 | 55.24661 | -3.227543 | 6.27184 |
b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg
Test: Ho: difference in coefficients not systematic
chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 41.89
Prob>chi2 = 0.0000
(V_b-V_B is not positive definite)
Kiểm định đa cộng tuyến
Collinearity Diagnostics
SQRT R-
Variable VIF VIF Tolerance Squared
---------------------------------------------------- lnsize 2.02 1.42 0.4938 0.5062
nim 1.20 1.10 0.830 8 0.1692
lta 1.14 1.07 0.8748 0.1252
eta 1.71 1.31 0.5857 0.4143
cir 1.04 1.02 0.9633 0.0367
gdp 1.30 1.14 0.7676 0.2324
---------------------------------------------------- Mean VIF 1.40
Cond
Eigenval Index
---- -----------------------------
5.0552 | 1.0000 | |
2 | 0.9893 | 2.2604 |
3 | 0.5302 | 3.0878 |
4 | 0.2645 | 4.3721 |
5 | 0.1334 | 6.1564 |
6 | 0.0256 | 14.0543 |
7 | 0.0019 | 51.9155 |
---------------------------------
Condition Number 51.9155
Eigenvalues & Cond Index computed from scaled raw sscp (w/ intercept) Det(correlation matrix) 0.3712
.
Modified Wald test for groupwise heteroskedasticity in fixed effect regression model
H0: sigma(i)^2 = sigma^2 for all i
chi2 (40) = 5.1e+30
Prob>chi2 = 0.9650
.
Kiểm định tự tương quan của phần dư
. xtserial sdroa lnsize nim lta eta cir gdp
Wooldridge test for autocorrelation in panel data H0: no first-order autocorrelation
F( 1, 38) = 1.375
Prob > F = 0.2482
.