Re: [Gretl-users] non-linear Wald-Test

Follow-up in regard to Jack's nice presentation of the delta method for testing nonlinear restrictions. This is not documented yet (nor finalized, if people have good suggestions for changes) but in CVS we have the following option for the "restrict" command: restrict rfunc = <function-name> end restrict That is, you can use the keyword "rfunc" to indicate that you're giving a constraint function rather than explicit statement of a restriction. As in Jack's script, this function should take as its single argument the coefficient vector (in "pointer" form), and it should return a vector that is zero under H0, non-zero under the alternative. At present this is available only for single-equation models. And the "rfunc" option cannot be combined with restrictions given in any other form within the same "restrict" block. Here's an example, in the neighborhood of Arthur T's question, testing whether two pairs of parameter values have a common ratio: <script> function restr (matrix *b) matrix ret = b[1]/b[2] - b[4]/b[5] return matrix ret end function open data4-10 ols 1 0 2 3 4 5 restrict rfunc = restr end restrict </script>