xtabond2 npl l.npl l2.loan, gmm(l.npl loan,lag(2 5) collapse) nolevel small robust

Favoring speed over space. To switch, type or click on mata: mata set matafavor space, perm.

Dynamic panel-data estimation, one-step difference GMM

------------------------------------------------------------------------------ Group variable: bank_dum Number of obs = 328 Time variable : year Number of groups = 48 Number of instruments = 8 Obs per group: min = 0 F(2, 48) = 16.93 avg = 6.83 Prob > F = 0.000 max = 10 ------------------------------------------------------------------------------ | Robust

npl | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- npl |

L1. | .9803696 .1714739 5.72 0.000 .6355982 1.325141 |

loan |

L2. | -.0167082 .0098918 -1.69 0.098 -.0365969 .0031806 ------------------------------------------------------------------------------ Instruments for first differences equation

GMM-type (missing=0, separate instruments for each period unless collapsed)

L(2/5).(L.npl loan) collapsed

------------------------------------------------------------------------------ Arellano-Bond test for AR(1) in first differences: z = -1.72 Pr > z = 0.085 Arellano-Bond test for AR(2) in first differences: z = -0.30 Pr > z = 0.763 ------------------------------------------------------------------------------ Sargan test of overid. restrictions: chi2(6) = 2.26 Prob > chi2 = 0.895 (Not robust, but not weakened by many instruments.)

Hansen test of overid. restrictions: chi2(6) = 5.79 Prob > chi2 = 0.447 (Robust, but weakened by many instruments.)

加入了控制变量size,结果变成了

xtabond2 npl l.npl l2.loan size, gmm(l.npl loan,lag(2 5) collapse) iv(size) nolevel small robust

Favoring speed over space. To switch, type or click on mata: mata set

matafavor space, perm.

Dynamic panel-data estimation, one-step difference GMM ------------------------------------------------------------------------------

Group variable: bank_dum Number of obs = 328 Time variable : year Number of groups = 48 Number of instruments = 9 Obs per group: min = 0 F(3, 48) = 10.82 avg = 6.83

Prob > F = 0.000 max = 10 ------------------------------------------------------------------------------ | Robust

npl | Coef. Std. Err. t P>|t| [95% Conf. Interval]

-------------+---------------------------------------------------------------- npl |

L1. | 1.018233 .1890329 5.39 0.000 .6381568 1.398309 |

loan |

L2. | -.0131397 .0101796 -1.29 0.203 -.0336071 .0073277 |

size | .3450704 .2908862 1.19 0.241 -.2397954 .9299363 Instruments for first differences equation

Standard

D.size

GMM-type (missing=0, separate instruments for each period unless collapsed)

L(2/5).(L.npl loan) collapsed

------------------------------------------------------------------------------ Arellano-Bond test for AR(1) in first differences: z = -1.76 Pr > z = 0.079 Arellano-Bond test for AR(2) in first differences: z = -0.29 Pr > z = 0.772 ------------------------------------------------------------------------------ Sargan test of overid. restrictions: chi2(6) = 3.28 Prob > chi2 = 0.773 (Not robust, but not weakened by many instruments.)

Hansen test of overid. restrictions: chi2(6) = 7.91 Prob > chi2 = 0.245 (Robust, but weakened by many instruments.)

注意：

首先，gmm（）里不能有lag的形式，所以不是gmm(l.npl loan,lag(2 5) collapse)，而是 gmm(npl loan,lag(2 5) collapse)；

其次，iv(size)其实你就默认size是严格外生的，不良贷款率也有可能影响size啊，

所以也有可能是内生；

再次，解释变量为什么选用滞后两期的 l2.loan，可以直接使用当期的loan试试； 最后，面板数据最好加入时间虚拟变量/行业虚拟变量/地区虚拟变量 等考虑。