请教一个C-D函数的问题

2341

收藏 2007-06-18

我用C- D 生产函数Y=A₀e^rtK^αL^β计算科技贡献率。为了消除多重共线性，用α+β=1去掉一个自变量，得㏑(Y/L) =㏑A₀+rt+α㏑(K/L)+u₂

然后剔除二阶自相关，得到的结果是自变量ln(k/l)的t统计无法通过，该怎么办？（不存在异方差）

Dependent Variable: LNGDP/L
Method: Least Squares
Date: 05/18/07 Time: 22:54
Sample (adjusted): 3 21
Included observations: 19 after adjustments
Convergence achieved after 11 iterations


Variable	Coefficient	Std. Error	t-Statistic	Prob.


LNK/L	0.175196	0.136118	1.287093	0.2189
T	0.109657	0.030037	3.650776	0.0026
C	-1.629247	0.591887	-2.752631	0.0156
AR(1)	1.490989	0.185746	8.027027	0.0000
AR(2)	-0.714602	0.195862	-3.648491	0.0026


R-squared	0.996879	Mean dependent var		-0.626140
Adjusted R-squared	0.995987	S.D. dependent var		0.857615
S.E. of regression	0.054326	Akaike info criterion		-2.766691
Sum squared resid	0.041318	Schwarz criterion		-2.518155
Log likelihood	31.28357	F-statistic		1117.953
Durbin-Watson stat	1.767394	Prob(F-statistic)		0.000000


Inverted AR Roots	.75-.40i	.75+.40i