其中需要使用iterative weighted nonlinear least squares估计

Has anybody know?pls help

你的问题解决了不？我现在也碰见了同样的问题。急需求救。

但是我在ssrn上看见一篇working paper“Regression-Based Tests of the Market Pricing of Accounting Numbers: The Mishkin Test and Ordinary Least Squares ”。他说mishkin，就可以直接使用ols

Program syntax (square brackets denote optional parameters):
*
*   mishkin varname1 varname2 [if ...] [in...], predvars(varlist)
*     [labwd(#)] [colwd(#)] [coldg(#)] [predict] [het]
*
* where:
*     varname1  Returns variable
*     varname2  Earnings variable
*     varlist   List of earnings predictor variables
*     labwd     Number indicating width of variable label column
*     colwd     Nmber indicating width of numeric column
*     het       Tells program to reweight forecast equation so that the returns
*               and forecast equations have the same root mean squared error
*
* Example usage:
* Estimate system where ret denotes current year returns, earn denotes current
* year earnings, accr denotes prior year accruals and cash denotes prior year cash
* flow from operations.  After estimation, generate predicted values of earnings based
* on the forecast equation (predF) and the return equation (predR).
*
*   mishkin ret earn if year==1995, predvars(accr cash) labwd(12) colwd(10) coldg(4)
*   matrix betaF=e(b_fcst)
*   matrix score predF=betaF
*   matrix betaR=e(b_retn)
*   matrix score predR=betaR

[此贴子已经被作者于2008-10-3 22:39:31编辑过]

詳細資料請參考http://personal.anderson.ucla.edu/judson.caskey/data.html