rls
Recursive least squares regression.
The rls view of an equation displays the results of recursive least squares (rolling) regression. This view is only available for (non-panel) equations estimated by ordinary least squares without ARMA terms.
You may plot various statistics from rls by choosing an option.
Syntax
eq_name.rls(options) c(1) c(2) …
Options

 r Plot the recursive residuals about the zero line with plus and minus two standard errors. r,s Plot the recursive residuals and save the residual series and their standard errors as series named R_RES and R_RESSE, respectively. c Plot the recursive coefficient estimates with two standard error bands. c,s Plot the listed recursive coefficients and save all coefficients and their standard errors as series named R_C1, R_C1SE, R_C2, R_C2SE, and so on. o Plot the p-values of recursive one-step Chow forecast tests. n Plot the p-values of recursive n-step Chow forecast tests. q Plot the CUSUM (standardized cumulative recursive residual) and 5 percent critical lines. v Plot the CUSUMSQ (CUSUM of squares) statistic and 5 percent critical lines. prompt Force the dialog to appear from within a program. p Print the view.
Examples
equation eq1.ls m1 c tb3 gdp
eq1.rls(r,s)
eq1.rls(c) c(2) c(3)
plots and saves the recursive residual series from EQ1 and their standard errors as R_RES and R_RESSE. The third line plots the recursive slope coefficients of EQ1.
equation eq2.ls m1 c pdl(tb3,12,3) pdl(gdp,12,3)
eq2.rls(c) c(3)
eq2.rls(q)
The second command plots the recursive coefficient estimates of PDL02, the linear term in the polynomial of TB3 coefficients. The third line plots the CUSUM test statistic and the 5% critical lines.
Cross-references