*! spregrext V2.0 03/03/2013 *! *! Emad Abd Elmessih Shehata *! Professor (PhD Economics) *! Agricultural Research Center - Agricultural Economics Research Institute - Egypt *! Email: emadstat@hotmail.com *! WebPage: http://emadstat.110mb.com/stata.htm *! WebPage at IDEAS: http://ideas.repec.org/f/psh494.html *! WebPage at EconPapers: http://econpapers.repec.org/RAS/psh494.htm *! Sahra Khaleel A. Mickaiel *! Professor (PhD Economics) *! Cairo University - Faculty of Agriculture - Department of Economics - Egypt *! Email: sahra_atta@hotmail.com *! WebPage: http://sahraecon.110mb.com/stata.htm *! WebPage at IDEAS: http://ideas.repec.org/f/pmi520.html *! WebPage at EconPapers: http://econpapers.repec.org/RAS/pmi520.htm program define spregrext, eclass version 11.0 syntax varlist , nc(str) Model(str) [WMFile(str) diag aux(str) INV INV2 /// vce(passthru) stand MFX(str) LMHet LMSPac zero level(passthru) /// coll PREDict(str) RESid(str) tolog TESTs DN NOCONStant LMNorm] gettoken yvar xvar : varlist local sthlp spregrext local both : list yvar & xvar if "`both'" != "" { di di as err " {bf:{cmd:`both'} included in both LHS and RHS Variables}" di as res " LHS: `yvar'" di as res " RHS: `xvar'" exit } if "`xvar'"=="" { di as err " {bf:Independent Variable(s) must be combined with Dependent Variable}" exit } local both : list xvar & aux if "`both'" != "" { di di as err " {bf:{cmd:`both'} included in both RHS and Auxiliary Variables}" di as res " RHS: `xvar'" di as res " AUX: `aux'" exit } if "`model'"!="" { if !inlist("`model'", "sar", "sdm") { di di as err "{bf:model( )} {cmd:options:} {bf:sar, sdm}" di in smcl _c "{cmd: see:} {help `sthlp'##03:Model Options}" di in gr _c " (spregrext Help):" exit } } if "`tests'"!="" { local lmspac "lmspac" local diag "diag" local lmhet "lmhet" local lmnorm "lmnorm" } if "`mfx'"!="" { if !inlist("`mfx'", "lin", "log") { di di as err " {bf:mfx( )} {cmd:must be} {bf:mfx({it:lin})} {cmd:for Linear Model, or} {bf:mfx({it:log})} {cmd:for Log-Log Model}" exit } } if inlist("`mfx'", "log") { if "`tolog'"=="" { di di as err " {bf:tolog} {cmd:must be combined with} {bf:mfx(log)}" exit } } tempvar D DE DF1 DumE DW E dcs XQX_ EE Hat ht idv itv LE LEo P Q Sig2 SSE tempvar SST Time U U2 Ue wald waldm waldp weit Wi Wio WS X X0 Bo tempvar Xb XB Xo XQ Yb Yh Yh2 Yhb Yho Yho2 Yt YY YYm YYv Yh_ML Ue_ML Z tempname Xb A B b Beta BetaSP Bm Bx Cov D den DVE DVNE Dx E E1 tempname J K M mh P Phi Pm Q In WS D WW eVec Xo kbm SSE SST1 SST2 tempname Sig2 Sig2b Sig2o Sig2u We Wi Wi1 Wio WY X X0 XB V F h Hat tempname Vec vh VM VP Vs W W1 W2 Wald Y Yh YYm YYv Z Dim kb IPhi tempname N DF kx kb NC NT Nmiss llf wmat xAx xAx2 TRa2 wWw1 xBx X0 Sig21 di scalar `N'=_N scalar `NC'=`nc' scalar `NT'=int(_N/`NC') local N=`N' local NT=`NT' local NC=`NC' if "`wmfile'" != "" { preserve qui use `"`wmfile'"', clear qui summ if `NT' !=int(`N'/r(N)) { di di as err " Cross Section Weight Matrix = " r(N) di as err " Time Series obs = " `NT' di as err " Sample Size (Number of obs) = " `N' di as res " {bf:(Cross Sections x Time) must be Equal Sample Size}" di as err " {bf:Spatial Cross Section Weight Matrix Dimension not correct}" di as err " {bf:Check Correct Number Units, Unequal Time Series not Allowed}" exit } mkmat * , matrix(_WB) qui egen ROWSUM=rowtotal(*) qui count if ROWSUM==0 local NN=r(N) if `NN'==1 { di as err "*** {bf:Spatial Weight Matrix Has (`NN') Location with No Neighbors}" } else if `NN'>1 { di as err "*** {bf:Spatial Weight Matrix Has (`NN') Locations with No Neighbors}" } local NROW=rowsof(_WB) local NCOL=colsof(_WB) if `NROW'!=`NCOL' { di as err "*** {bf:Spatial Weight Matrix is not Square}" exit } di _dup(78) "{bf:{err:=}}" if "`stand'"!="" { di as res "{bf:*** Standardized Weight Matrix: `N'x`N' - NC=`NC' NT=`NT' (Normalized)}" matrix `Xo'=J(`NC',1,1) matrix WB=_WB*`Xo'*`Xo'' mata: X = st_matrix("_WB") mata: Y = st_matrix("WB") mata: _WS=X:/Y mata: _WS=st_matrix("_WS",_WS) mata: _WS = st_matrix("_WS") if "`inv'"!="" { di as res "{bf:*** Inverse Standardized Weight Matrix (1/W)}" mata: _WS=1:/_WS mata: _editmissing(_WS, 0) mata: _WS=st_matrix("_WS",_WS) } if "`inv2'"!="" { di as res "{bf:*** Inverse Squared Standardized Weight Matrix (1/W^2)}" mata: _WS=_WS:*_WS mata: _WS=1:/_WS mata: _editmissing(_WS, 0) mata: _WS=st_matrix("_WS",_WS) } matrix WCS=_WS } else { di as res "{bf:*** Binary (0/1) Weight Matrix: `N'x`N' - NC=`NC' NT=`NT' (Non Normalized)}" matrix WCS=_WB } matrix WMB=WCS#I(`NT') di _dup(78) "{bf:{err:=}}" restore } ereturn scalar Nn=_N qui marksample touse qui cap count local N = r(N) qui count local N = _N qui markout `touse' `varlist' , strok qui mean `varlist' scalar `Nmiss' = e(N) if "`zero'"=="" { if `N' !=`Nmiss' { di di as err "*** {bf:Observations have {bf:(" `N'-`Nmiss' ")} Missing Values}" di as err "*** {bf:You can use {cmd:zero} option to Convert Missing Values to Zero}" exit } } qui cap count local N = r(N) local NC=`nc' local NT=`N'/`NC' qui cap drop `idv' qui cap drop `itv' qui cap drop `Time' gen `Time' = _n gen `idv'= ceil(_n/`NT') gen `itv'= _n-(`idv'-1)*`NT' qui xtset `idv' `itv' scalar `Dim' = `N' local MSize= `Dim' if `c(matsize)' < `MSize' { di as err " {bf:Current Matrix Size = (`c(matsize)')}" di as err " {bf:{help matsize##|_new:matsize} must be >= Sample Size" as res " (`MSize')}" qui set matsize `MSize' di as res " {bf:matsize increased now to = (`c(matsize)')}" } qui marksample touse markout `touse' `varlist' , strok local NC=`NC' local NT=`NT' if "`zero'"!="" { tempvar zeromiss qui foreach var of local varlist { qui gen `zeromiss'`var'=`var' qui replace `var'=0 if `var'==. } } _vce_parse , opt(Robust oim opg) argopt(CLuster): , `vce' if "`tolog'"!="" { local vlistlog " `varlist' `mhet' " qui _rmcoll `vlistlog' , `noconstant' `coll' forcedrop local vlistlog "`r(varlist)'" di _dup(78) "-" di as err " {cmd:** Data Have been Transformed to Log Form **}" di as txt " {cmd:** `varlist'} " di _dup(78) "-" qui foreach var of local vlistlog { tempvar xyind`var' qui gen `xyind`var''=`var' qui replace `var'=ln(`var') qui replace `var'=0 if `var'==. } } if "`coll'"=="" { _rmcoll `varlist' , `noconstant' `coll' forcedrop local varlist "`r(varlist)'" gettoken yvar xvar : varlist if "`aux'"!="" { _rmcoll `aux' , `noconstant' `coll' forcedrop local aux "`r(varlist)'" } } qui cap drop w1x_* qui cap drop w1y_* mkmat `yvar' , matrix(`Y') tempname WS1 xyvar mkmat `yvar' , matrix(`xyvar') matrix `WS1'= WMB qui cap drop w1x_* qui cap drop w1y_* tempname w1y_`yvar' matrix `w1y_`yvar'' = `WS1'*`xyvar' svmat `w1y_`yvar'' , name(w1y_`yvar') rename w1y_`yvar'1 w1y_`yvar' label variable w1y_`yvar' `"AR(1) `yvar' Spatial Lag"' qui foreach var of local xvar { qui cap drop w1x_`var' tempname w1x_`var' mkmat `var' , matrix(`xyvar') matrix `w1x_`var'' = `WS1'*`xyvar' qui cap drop w1x_`var' svmat `w1x_`var'' , name(w1x_`var') rename w1x_`var'1 w1x_`var' label variable w1x_`var' `"AR(1) `var' Spatial Lag"' } qui gen `X0'=1 qui mkmat `X0' , matrix(`X0') qui tabulate `idv' , generate(`dcs') mkmat `dcs'* , matrix(`D') matrix `P'=`D'*invsym(`D''*`D')*`D'' matrix `Q'=I(`N')-`P' if inlist("`model'", "sar") { unab zvar: w1y_`yvar' local SPXvar `zvar' `xvar' `aux' } if inlist("`model'", "sdm") { unab zvar: w1x_* local SPXvar `xvar' `zvar' `aux' } if "`coll'"=="" { _rmcoll `SPXvar' , `noconstant' `coll' forcedrop local SPXvar "`r(varlist)'" } di _dup(78) "{bf:{err:=}}" if inlist("`model'", "sar") { di as txt "{bf:{err:* Spatial Panel Random-Effects Lag Regression (SAR)}}" } if inlist("`model'", "sdm") { di as txt "{bf:{err:* Spatial Panel Random-Effects Durbin Regression (SDM)}}" } di _dup(78) "{bf:{err:=}}" local kx : word count `SPXvar' local kaux : word count `aux' scalar `kb'=`kx'+1 scalar `DF'=`N'-`kx'-`NC' if "`noconstant'"!="" { qui mkmat `SPXvar' , matrix(`X') scalar `kb'=`kx' qui mean `SPXvar' } else { qui mkmat `SPXvar' `X0' , matrix(`X') scalar `kb'=`kx'+1 qui mean `SPXvar' `X0' } matrix `Xb'=e(b) qui mean `yvar' matrix `Yb'=e(b)' if "`dn'"!="" { scalar `DF'=`N' } global spat_kx=`kx' global spat_kxw=2*`kx' ereturn scalar df_m=$spat_kx qui gen `Wi'=1 qui gen `weit' = 1 mkmat `Wi' , matrix(`Wi') matrix `Wi'=diag(`Wi') scalar `llf'=. matrix `In'=I(`N') qui xtset `idv' `itv' qui xtreg `yvar' `SPXvar' , mle `vce' `level' `noconstant' scalar `llf'=e(ll) matrix `Beta'=e(b) matrix `Cov'= e(V) matrix `Beta'=`Beta'[1,1..`kb']' matrix `E'=`Y'-`X'*`Beta' matrix `Sig2'=`E''*`E'/`DF' matrix `Cov' = `Cov'[1..`kb', 1..`kb'] ereturn scalar df_m=$spat_kx if inlist("`model'", "sar") { qui test _b[w1y_`yvar'] =0 tempname rRo fgs2s rdfgs rdfgsr fgs2sp scalar `rRo'=_b[w1y_`yvar'] scalar `fgs2s'=r(chi2) scalar `rdfgs'=r(df) scalar `fgs2sp'=r(p) scalar `rRo'=`Beta'[1,1] } matrix `Yh_ML'=`X'*`Beta' matrix `Ue_ML'=`Q'*(`Y'-`Yh_ML') qui svmat `Yh_ML' , name(`Yh_ML') qui rename `Yh_ML'1 `Yh_ML' qui svmat `Ue_ML' , name(`Ue_ML') qui rename `Ue_ML'1 `Ue_ML' local N=_N if "`predict'"!= "" { qui cap drop `predict' qui gen `predict'=`Yh_ML' label variable `predict' `"Yh - Prediction"' } if "`resid'"!= "" { qui cap drop `resid' qui gen `resid'=`Ue_ML' label variable `resid' `"Ue - Residual"' } tempname SSEo Sigo r2bu r2bu_a r2raw r2raw_a R20 f fp wald waldp Sig2n tempname r2v r2v_a fv fvp r2h r2h_a fh fhp SSTm SSE1 SST11 SST21 Rho matrix `SSE'=`Ue_ML''*`Ue_ML' scalar `SSEo'=`SSE'[1,1] scalar `Sig2o'=`SSEo'/`DF' scalar `Sig2'=`SSEo'/`DF' scalar `Sig2n'=`SSEo'/`N' scalar `Sigo'=sqrt(`Sig2o') qui summ `Yh_ML' local NUM=r(Var) qui summ `yvar' local DEN=r(Var) scalar `r2v'=`NUM'/`DEN' scalar `r2v_a'=1-((1-`r2v')*(`N'-1)/`DF') scalar `fv'=`r2v'/(1-`r2v')*(`N'-`kb')/(`kx') scalar `fvp'=Ftail(`kx', `DF', `fv') qui correlate `Yh_ML' `yvar' scalar `r2h'=r(rho)*r(rho) scalar `r2h_a'=1-((1-`r2h')*(`N'-1)/`DF') scalar `fh'=`r2h'/(1-`r2h')*(`N'-`kb')/(`kx') scalar `fhp'=Ftail(`kx', `DF', `fh') qui summ `yvar' local Yb=r(mean) qui gen `YYm'=(`yvar'-`Yb')^2 qui summ `YYm' qui scalar `SSTm' = r(sum) qui gen `YYv'=(`yvar')^2 qui summ `YYv' local SSTv = r(sum) qui summ `weit' qui gen `Wi1'=sqrt(`weit'/r(mean)) mkmat `Wi1' , matrix(`Wi1') matrix `P' =diag(`Wi1') qui gen `Wio'=(`Wi1') mkmat `Wio' , matrix(`Wio') matrix `Wio'=diag(`Wio') matrix `Pm' =`Wio' matrix `IPhi'=`P''*`P' matrix `Phi'=invsym(`P''*`P') matrix `J'= J(`N',1,1) matrix `D'=(`J'*`J''*`IPhi')/`N' matrix `SSE'=`Ue_ML''*`IPhi'*`Ue_ML' matrix `SST1'=(`Y'-`D'*`Y')'*`IPhi'*(`Y'-`D'*`Y') matrix `SST2'=(`Y''*`Y') scalar `SSE1'=`SSE'[1,1] scalar `SST11'=`SST1'[1,1] scalar `SST21'=`SST2'[1,1] scalar `r2bu'=1-`SSE1'/`SST11' if `r2bu'< 0 { scalar `r2bu'=`r2h' } scalar `r2bu_a'=1-((1-`r2bu')*(`N'-1)/`DF') scalar `r2raw'=1-`SSE1'/`SST21' scalar `r2raw_a'=1-((1-`r2raw')*(`N'-1)/`DF') scalar `f'=`r2bu'/(1-`r2bu')*(`N'-`kb')/`kx' scalar `fp'= Ftail(`kx', `DF', `f') scalar `wald'=`f'*`kx' scalar `waldp'=chi2tail(`kx', abs(`wald')) if `llf' == . { scalar `llf'=-(`N'/2)*log(2*_pi*`Sig2n')-(`N'/2) } local Nof =`N' local Dof =`DF' matrix `B'=`Beta'' if "`noconstant'"!="" { matrix colnames `Cov' = `SPXvar' matrix rownames `Cov' = `SPXvar' matrix colnames `B' = `SPXvar' } else { matrix colnames `Cov' = `SPXvar' _cons matrix rownames `Cov' = `SPXvar' _cons matrix colnames `B' = `SPXvar' _cons } yxregeq `yvar' `SPXvar' di as txt _col(3) "Sample Size" _col(21) "=" %12.0f as res `N' _col(37) "|" _col(41) as txt "Cross Sections Number" _col(65) "=" _col(73) %5.0f as res `nc' ereturn post `B' `Cov' , dep(`yvar') obs(`Nof') dof(`Dof') qui test `SPXvar' scalar `f'=r(F) scalar `fp'= Ftail(`kx', `DF', `f') scalar `wald'=`f'*`kx' scalar `waldp'=chi2tail(`kx', abs(`wald')) di as txt _col(3) "{cmd:Wald Test}" _col(21) "=" %12.4f as res `wald' _col(37) "|" _col(41) as txt "P-Value > {cmd:Chi2}(" as res `kx' ")" _col(65) "=" %12.4f as res `waldp' di as txt _col(3) "{cmd:F-Test}" _col(21) "=" %12.4f as res `f' _col(37) "|" _col(41) as txt "P-Value > {cmd:F}(" as res `kx' " , " `DF' ")" _col(65) "=" %12.4f as res `fp' di as txt _col(2) "(Buse 1973) R2" _col(21) "=" %12.4f as res `r2bu' _col(37) "|" as txt _col(41) "Raw Moments R2" _col(65) "=" %12.4f as res `r2raw' ereturn scalar r2bu =`r2bu' ereturn scalar r2bu_a=`r2bu_a' ereturn scalar f =`f' ereturn scalar fp=`fp' ereturn scalar wald =`wald' ereturn scalar waldp=`waldp' di as txt _col(2) "(Buse 1973) R2 Adj" _col(21) "=" %12.4f as res `r2bu_a' _col(37) "|" as txt _col(41) "Raw Moments R2 Adj" _col(65) "=" %12.4f as res `r2raw_a' di as txt _col(3) "Root MSE (Sigma)" _col(21) "=" %12.4f as res `Sigo' as txt _col(37) "|" _col(41) "Log Likelihood Function" _col(65) "=" %12.4f as res `llf' di _dup(78) "-" di as txt "- {cmd:R2h}=" %7.4f as res `r2h' _col(17) as txt "{cmd:R2h Adj}=" as res %7.4f `r2h_a' as txt _col(34) "{cmd:F-Test} =" %8.2f as res `fh' _col(51) as txt "P-Value > F(" as res `kx' " , " `DF' ")" _col(72) %5.4f as res `fhp' if `r2v'<1 { di as txt "- {cmd:R2v}=" %7.4f as res `r2v' _col(17) as txt "{cmd:R2v Adj}=" as res %7.4f `r2v_a' as txt _col(34) "{cmd:F-Test} =" %8.2f as res `fv' _col(51) as txt "P-Value > F(" as res `kx' " , " `DF' ")" _col(72) %5.4f as res `fvp' ereturn scalar r2v=`r2v' ereturn scalar r2v_a=`r2v_a' ereturn scalar fv=`fv' ereturn scalar fvp=`fvp' } ereturn scalar r2raw =`r2raw' ereturn scalar r2raw_a=`r2raw_a' ereturn scalar llf =`llf' ereturn scalar sig=`Sigo' ereturn scalar r2h =`r2h' ereturn scalar r2h_a=`r2h_a' ereturn scalar fh =`fh' ereturn scalar fhp=`fhp' ereturn scalar kb=`kb' ereturn scalar kx=`kx' ereturn scalar DF=`DF' ereturn scalar Nn=_N ereturn scalar NC=`NC' ereturn scalar NT=`NT' ereturn display , `level' matrix `b'=e(b) matrix `V'=e(V) matrix `Bx' =`b'[1..1, 1..`kx']' if inlist("`model'", "sar") { qui test _b[w1y_`yvar'] =0 tempname rRo fgs2s rdfgs rdfgsr fgs2sp scalar `rRo'=_b[w1y_`yvar'] scalar `fgs2s'=r(F) scalar `rdfgs'=r(df) scalar `fgs2sp'=r(p) di as txt _col(3) "Rho Value =" %8.4f as res `rRo' _col(30) as txt "Chi2 Test =" %10.3f as res `fgs2s' as txt _col(54) "P-Value > Chi2(1)" _col(73) %5.4f as res `fgs2sp' di _dup(78) "-" } matrix `b'=e(b) matrix `V'=e(V) matrix `Bx' =`b'[1..1, 1..`kx']' if "`diag'"!= "" { di di _dup(78) "{bf:{err:=}}" di as txt "{bf:{err:* Panel Model Selection Diagnostic Criteria}}" di _dup(78) "{bf:{err:=}}" scalar `kbm'=`kb'+3 ereturn scalar aic=`Sig2n'*exp(2*`kbm'/`N') ereturn scalar laic=ln(`Sig2n')+2*`kbm'/`N' ereturn scalar fpe=`Sig2o'*(1+`kbm'/`N') ereturn scalar sc=`Sig2n'*`N'^(`kbm'/`N') ereturn scalar lsc=ln(`Sig2n')+`kbm'*ln(`N')/`N' ereturn scalar hq=`Sig2n'*ln(`N')^(2*`kbm'/`N') ereturn scalar rice=`Sig2n'/(1-2*`kbm'/`N') ereturn scalar shibata=`Sig2n'*(`N'+2*`kbm')/`N' ereturn scalar gcv=`Sig2n'*(1-`kbm'/`N')^(-2) ereturn scalar llf = `llf' di as txt "- Log Likelihood Function" _col(45) "LLF" _col(60) "=" %12.4f `e(llf)' di _dup(75) "-" di as txt "- Akaike Information Criterion" _col(45) "(1974) AIC" _col(60) "=" %12.4f `e(aic)' di as txt "- Akaike Information Criterion" _col(45) "(1973) Log AIC" _col(60) "=" %12.4f `e(laic)' di _dup(75) "-" di as txt "- Schwarz Criterion" _col(45) "(1978) SC" _col(60) "=" %12.4f `e(sc)' di as txt "- Schwarz Criterion" _col(45) "(1978) Log SC" _col(60) "=" %12.4f `e(lsc)' di _dup(75) "-" di as txt "- Amemiya Prediction Criterion" _col(45) "(1969) FPE" _col(60) "=" %12.4f `e(fpe)' di as txt "- Hannan-Quinn Criterion" _col(45) "(1979) HQ" _col(60) "=" %12.4f `e(hq)' di as txt "- Rice Criterion" _col(45) "(1984) Rice" _col(60) "=" %12.4f `e(rice)' di as txt "- Shibata Criterion" _col(45) "(1981) Shibata" _col(60) "=" %12.4f `e(shibata)' di as txt "- Craven-Wahba Generalized Cross Validation" _col(45) "(1979) GCV" _col(60) "=" %12.4f `e(gcv)' di _dup(78) "-" } if "`lmspac'"!= "" { tempname B0 B1 B2 b2k B3 B4 chi21 chi22 DEN DIM E EI Ein Esp eWe tempname I m1k m2k m3k m4k MEAN NUM NUM1 NUM2 NUM3 RJ WXb IN xMx tempname S0 S1 S2 sd SEI SG0 SG1 TRa1 TRa2 trb TRw2 VI wi wj wMw wMw1 tempname WZ0 Xk NUM Zk m2 m4 b2 M eWe1 wWw2 zWz eWy eWy1 CPX SUM DFr tempvar WZ0 Vm2 Vm4 matrix `wmat'=_WB#I(`NT') scalar `DFr'=`N'-`kb' scalar `S1'=0 qui forvalue i = 1/`N' { qui forvalue j = 1/`N' { scalar `S1'=`S1'+(`wmat'[`i',`j']+`wmat'[`j',`i'])^2 local j=`j'+1 } } matrix `zWz'=`X0''*`wmat'*`X0' scalar `S0'=`zWz'[1,1] scalar `SG0'=`S0'*2 matrix `WZ0' =`wmat'*`X0' qui svmat `WZ0' , name(`WZ0') qui rename `WZ0'1 `WZ0' qui replace `WZ0'=(`WZ0'+`WZ0')^2 qui summ `WZ0' scalar `SG1'=r(sum) matrix `E'=`Ue_ML' matrix `eWe'=`E''*`wmat'*`E' scalar `eWe1'=`eWe'[1,1] matrix `CPX'=`X''*`X' matrix `A'=inv(`CPX') matrix `xAx'=`A'*`X''*`wmat'*`X' scalar `TRa1'=trace(`xAx') matrix `xAx2'=`xAx'*`xAx' scalar `TRa2'=trace(`xAx2') matrix `wWw1'=(`wmat'+`wmat'')*(`wmat'+`wmat'') matrix `B'=inv(`CPX') matrix `xBx'=`B'*`X''*`wWw1'*`X' scalar `trb'=trace(`xBx') scalar `VI'=(`N'^2/(`N'^2*`DFr'*(2+`DFr')))*((`S1'/2)+2*`TRa2'-`trb'-2*`TRa1'^2/`DFr') scalar `SEI'=sqrt(`VI') scalar `I'=(`N'/`S0')*`eWe1'/`SSEo' scalar `EI'=-(`N'*`TRa1')/(`DFr'*`N') ereturn scalar mi1=(`I'-`EI')/`SEI' ereturn scalar mi1p=2*(1-normal(abs(e(mi1)))) matrix `wWw2'=`wmat''*`wmat'+`wmat'*`wmat' scalar `TRw2'=trace(`wWw2') matrix `WY'=`wmat'*`Y' matrix `eWy'=`E''*`WY' scalar `eWy1'=`eWy'[1,1] matrix `WXb'=`wmat'*`X'*`Beta' matrix `IN'=I(`N') matrix `M'=inv(`CPX') matrix `xMx'=`IN'-`X'*`M'*`X'' matrix `wMw'=`WXb''*`xMx'*`WXb' scalar `wMw1'=`wMw'[1,1] scalar `RJ'=1/(`TRw2'+`wMw1'/`Sig2o') ereturn scalar lmerr=((`eWe1'/`Sig2o')^2)/`TRw2' ereturn scalar lmlag=((`eWy1'/`Sig2o')^2)/(`TRw2'+`wMw1'/`Sig2o') ereturn scalar lmerrr=(`eWe1'/`Sig2o'-`TRw2'*`RJ'*(`eWy1'/`Sig2o'))^2/(`TRw2'-`TRw2' *`TRw2'*`RJ') ereturn scalar lmlagr=(`eWy1'/`Sig2o'-`eWe1'/`Sig2o')^2/((1/`RJ')-`TRw2') ereturn scalar lmsac1=e(lmerr)+e(lmlagr) ereturn scalar lmsac2=e(lmlag)+e(lmerrr) ereturn scalar lmerrp=chi2tail(1, abs(e(lmerr))) ereturn scalar lmerrrp=chi2tail(1, abs(e(lmerrr))) ereturn scalar lmlagp=chi2tail(1, abs(e(lmlag))) ereturn scalar lmlagrp=chi2tail(1, abs(e(lmlagr))) ereturn scalar lmsac1p=chi2tail(2, abs(e(lmsac1))) ereturn scalar lmsac2p=chi2tail(2, abs(e(lmsac2))) scalar `DIM'=rowsof(`wmat') matrix `m2'=J(1,1,0) matrix `m4'=J(1,1,0) matrix `b2'=J(1,1,0) matrix `M'=J(1,4,0) local j=1 while `j'<=4 { tempvar EQsq qui gen `EQsq' = `Ue_ML'^`j' qui summ `EQsq' , mean matrix `M'[1,`j'] = r(sum) local j=`j'+1 } qui summ `Ue_ML' , mean scalar `MEAN'=r(mean) qui replace `Ue_ML'=`Ue_ML'-`MEAN' qui gen `Vm2'=`Ue_ML'^2 qui summ `Vm2' , mean matrix `m2'[1,1]=r(mean) scalar `m2k'=r(mean) qui gen `Vm4'=`Ue_ML'^4 qui summ `Vm4' , mean matrix `m4'[1,1]=r(mean) scalar `m4k'=r(mean) matrix `b2'[1,1]=`m4k'/(`m2k'^2) mkmat `Ue_ML' , matrix(`Ue_ML') scalar `S0'=0 scalar `S1'=0 scalar `S2'=0 local i=1 while `i'<=`N' { scalar `wi'=0 scalar `wj'=0 local j=1 while `j'<=`N' { scalar `S0'=`S0'+`wmat'[`i',`j'] scalar `S1'=`S1'+(`wmat'[`i',`j']+`wmat'[`j',`i'])^2 scalar `wi'=`wi'+`wmat'[`i',`j'] scalar `wj'=`wj'+`wmat'[`j',`i'] local j=`j'+1 } scalar `S2'=`S2'+(`wi'+`wj')^2 local i=`i'+1 } scalar `S1'=`S1'/2 scalar `m2k'=`m2'[1,1] scalar `b2k'=`b2'[1,1] matrix `Zk'=`Ue_ML'[1...,1] matrix `Zk'=`Zk''*`wmat'*`Zk' scalar `Ein'=-1/(`N'-1) scalar `NUM1'=`N'*((`N'^2-3*`N'+3)*`S1'-(`N'*`S2')+(3*`S0'^2)) scalar `NUM2'=`b2k'*((`N'^2-`N')*`S1'-(2*`N'*`S2')+(6*`S0'^2)) scalar `DEN'=(`N'-1)*(`N'-2)*(`N'-3)*(`S0'^2) scalar `sd'=sqrt((`NUM1'-`NUM2')/`DEN'-(1/(`N'-1))^2) ereturn scalar mig=`Zk'[1,1]/(`S0'*`m2k') ereturn scalar migz=(e(mig)-`Ein')/`sd' ereturn scalar migp=2*(1-normal(abs(e(migz)))) ereturn scalar mi1z=(e(mi1)-`Ein')/`sd' scalar `m2k'=`m2'[1,1] scalar `b2k'=`b2'[1,1] matrix `Zk'=`Ue_ML'[1...,1] scalar `SUM'=0 local i=1 while `i'<=`N' { local j=1 while `j'<=`N' { scalar `SUM'=`SUM'+`wmat'[`i',`j']*(`Zk'[`i',1]-`Zk'[`j',1])^2 local j=`j'+1 } local i=`i'+1 } scalar `NUM1'=(`N'-1)*`S1'*(`N'^2-3*`N'+3-(`N'-1)*`b2k') scalar `NUM2'=(1/4)*(`N'-1)*`S2'*(`N'^2+3*`N'-6-(`N'^2-`N'+2)*`b2k') scalar `NUM3'=(`S0'^2)*(`N'^2-3-((`N'-1)^2)*`b2k') scalar `DEN'=(`N')*(`N'-2)*(`N'-3)*(`S0'^2) scalar `sd'=sqrt((`NUM1'-`NUM2'+`NUM3')/`DEN') ereturn scalar gcg=((`N'-1)*`SUM')/(2*`N'*`S0'*`m2k') ereturn scalar gcgz=(e(gcg)-1)/`sd' ereturn scalar gcgp=2*(1-normal(abs(e(gcgz)))) scalar `B0'=((`N'^2)-3*`N'+3)*`S1'-`N'*`S2'+3*(`S0'^2) scalar `B1'=-(((`N'^2)-`N')*`S1'-2*`N'*`S2'+6*(`S0'^2)) scalar `B2'=-(2*`N'*`S1'-(`N'+3)*`S2'+6*(`S0'^2)) scalar `B3'=4*(`N'-1)*`S1'-2*(`N'+1)*`S2'+8*(`S0'^2) scalar `B4'=`S1'-`S2'+(`S0'^2) scalar `m1k'=`M'[1,1] scalar `m2k'=`M'[1,2] scalar `m3k'=`M'[1,3] scalar `m4k'=`M'[1,4] matrix `Xk'=`Ue_ML'[1...,1] matrix `NUM'=`Xk''*`wmat'*`Xk' scalar `DEN'=0 local i=1 while `i'<=`N' { local j=1 while `j'<=`N' { if `i'!=`j' { scalar `DEN'=`DEN'+`Xk'[`i',1]*`Xk'[`j',1] } local j=`j'+1 } local i=`i'+1 } ereturn scalar gog=`NUM'[1,1]/`DEN' scalar `Esp'=`S0'/(`N'*(`N'-1)) scalar `NUM'=(`B0'*`m2k'^2)+(`B1'*`m4k')+(`B2'*`m1k'^2*`m2k') /// +(`B3'*`m1k'*`m3k')+(`B4'*`m1k'^4) scalar `DEN'=(((`m1k'^2)-`m2k')^2)*`N'*(`N'-1)*(`N'-2)*(`N'-3) scalar `sd'=(`NUM'/`DEN')-((`Esp')^2) ereturn scalar gogz=(e(gog)-`Esp')/sqrt(`sd') ereturn scalar gogp=2*(1-normal(abs(e(gogz)))) scalar `chi21'=invchi2(1,0.95) scalar `chi22'=invchi2(2,0.95) di di _dup(78) "{bf:{err:=}}" di as txt "{bf:{err:*** Spatial Panel Aautocorrelation Tests}}" di _dup(78) "{bf:{err:=}}" di as txt _col(2) "{bf: Ho: Error has No Spatial AutoCorrelation}" di as txt _col(2) "{bf: Ha: Error has Spatial AutoCorrelation}" di di as txt "- GLOBAL Moran MI" as res _col(30) "=" %9.4f e(mig) _col(45) as txt "P-Value > Z(" %6.3f e(migz) ")" _col(67) %5.4f as res e(migp) di as txt "- GLOBAL Geary GC" as res _col(30) "=" %9.4f e(gcg) _col(45) as txt "P-Value > Z(" %5.3f e(gcgz) ")" _col(67) %5.4f as res e(gcgp) di as txt "- GLOBAL Getis-Ords GO" as res _col(30) "=" %9.4f e(gog) as txt _col(45) "P-Value > Z(" %5.3f e(gogz) ")" _col(67) %5.4f as res e(gogp) di _dup(78) "-" di as txt "- Moran MI Error Test" as res _col(30) "=" %9.4f e(mi1) _col(45) as txt "P-Value > Z(" %5.3f e(mi1z) ")" _col(67) %5.4f as res e(mi1p) di _dup(78) "-" di as txt "- LM Error (Burridge)" as res _col(30) "=" %9.4f e(lmerr) _col(45) as txt "P-Value > Chi2(1)" _col(67) %5.4f as res e(lmerrp) di as txt "- LM Error (Robust)" as res _col(30) "=" %9.4f e(lmerrr) _col(45) as txt "P-Value > Chi2(1)" _col(67) %5.4f as res e(lmerrrp) di _dup(78) "-" di as txt _col(2) "{bf: Ho: Spatial Lagged Dependent Variable has No Spatial AutoCorrelation}" di as txt _col(2) "{bf: Ha: Spatial Lagged Dependent Variable has Spatial AutoCorrelation}" di di as txt "- LM Lag (Anselin)" as res _col(30) "=" %9.4f e(lmlag) _col(45) as txt "P-Value > Chi2(1)" _col(67) %5.4f as res e(lmlagp) di as txt "- LM Lag (Robust)" as res _col(30) "=" %9.4f e(lmlagr) _col(45) as txt "P-Value > Chi2(1)" _col(67) %5.4f as res e(lmlagrp) di _dup(78) "-" di as txt _col(2) "{bf: Ho: No General Spatial AutoCorrelation}" di as txt _col(2) "{bf: Ha: General Spatial AutoCorrelation}" di di as txt "- LM SAC (LMErr+LMLag_R)" as res _col(30) "=" %9.4f e(lmsac2) _col(45) as txt "P-Value > Chi2(2)" _col(67) %5.4f as res e(lmsac2p) di as txt "- LM SAC (LMLag+LMErr_R)" as res _col(30) "=" %9.4f e(lmsac1) _col(45) as txt "P-Value > Chi2(2)" _col(67) %5.4f as res e(lmsac1p) di _dup(78) "-" } if "`lmhet'"!= "" { tempvar Yh Yh2 LYh2 E E2 E3 E4 LnE2 absE time LE ht DumE EDumE U2 tempname mh vh h Q LMh_cwx lmhmss1 mssdf1 lmhmss1p tempname lmhmss2 mssdf2 lmhmss2p dfw0 lmhw01 lmhw01p lmhw02 lmhw02p dfw1 lmhw11 tempname lmhw11p lmhw12 lmhw12p dfw2 lmhw21 lmhw21p lmhw22 lmhw22p lmhharv tempname lmhharvp lmhwald lmhwaldp lmhhp1 lmhhp1p lmhhp2 lmhhp2p lmhhp3 lmhhp3p lmhgl tempname lmhglp lmhcw1 cwdf1 lmhcw1p lmhcw2 cwdf2 lmhcw2p lmhq lmhqp qui tsset `Time' matrix `SSE'=`Ue_ML''*`Ue_ML' qui gen `U2' =`Ue_ML'^2/`Sig2n' qui gen `Yh'=`Yh_ML' qui gen `Yh2'=`Yh_ML'^2 qui gen `LYh2'=ln(`Yh2') qui gen `E' =`Ue_ML' qui gen `E2'=`Ue_ML'^2 qui gen `E3'=`Ue_ML'^3 qui gen `E4'=`Ue_ML'^4 qui gen `LnE2'=log(`E2') qui gen `absE'=abs(`E') qui gen `DumE'=0 qui replace `DumE'=1 if `E' >= 0 qui summ `DumE' qui gen `EDumE'=`E'*(`DumE'-r(mean)) qui regress `EDumE' `Yh' `Yh2' , `noconstant' `vce' scalar `lmhmss1'=e(N)*e(r2) scalar `mssdf1'=e(df_m) scalar `lmhmss1p'=chi2tail(`mssdf1', abs(`lmhmss1')) qui regress `EDumE' `SPXvar' , `noconstant' `vce' scalar `lmhmss2'=e(N)*e(r2) scalar `mssdf2'=e(df_m) scalar `lmhmss2p'=chi2tail(`mssdf2', abs(`lmhmss2')) local kx =`kx' qui forvalue j=1/`kx' { qui foreach i of local SPXvar { tempvar VN gen `VN'`j'=`i' qui cap drop `XQX_'`i' gen `XQX_'`i' = `VN'`j'*`VN'`j' } } qui regress `E2' `SPXvar' , `noconstant' scalar `dfw0'=e(df_m) scalar `lmhw01'=e(r2)*e(N) scalar `lmhw01p'=chi2tail(`dfw0' , abs(`lmhw01')) scalar `lmhw02'=e(mss)/(2*`Sig2n'^2) scalar `lmhw02p'=chi2tail(`dfw0' , abs(`lmhw02')) qui regress `E2' `SPXvar' `XQX_'* , `noconstant' scalar `dfw1'=e(df_m) scalar `lmhw11'=e(r2)*e(N) scalar `lmhw11p'=chi2tail(`dfw1' , abs(`lmhw11')) scalar `lmhw12'=e(mss)/(2*`Sig2n'^2) scalar `lmhw12p'=chi2tail(`dfw1' , abs(`lmhw12')) qui cap drop `VN'* qui cap drop `XQX_'* mkmat `SPXvar' , matrix(`VN') qui svmat `VN' , names(`VN') unab WWVN: `VN'* local ZWvar `WWVN' qui foreach i of local ZWvar { qui foreach j of local ZWvar { if `i' >= `j' { qui cap drop `XQX_'`i'`j' qui gen `XQX_'`i'`j' = `i'*`j' } } } qui cap matrix drop `VN' qui cap drop `VN'* qui regress `E2' `SPXvar' `XQX_'* , `noconstant' scalar `dfw2'=e(df_m) scalar `lmhw21'=e(r2)*e(N) scalar `lmhw21p'=chi2tail(`dfw2' , abs(`lmhw21')) scalar `lmhw22'=e(mss)/(2*`Sig2n'^2) scalar `lmhw22p'=chi2tail(`dfw2' , abs(`lmhw22')) qui regress `LnE2' `SPXvar' , `noconstant' scalar `lmhharv'=e(mss)/4.9348 scalar `lmhharvp'= chi2tail(2, abs(`lmhharv')) qui regress `LnE2' `SPXvar' , `noconstant' scalar `lmhwald'=e(mss)/2 scalar `lmhwaldp'=chi2tail(1, abs(`lmhwald')) qui regress `E2' `Yh' , `noconstant' scalar `lmhhp1'=e(N)*e(r2) scalar `lmhhp1p'=chi2tail(1, abs(`lmhhp1')) qui regress `E2' `Yh2' , `noconstant' scalar `lmhhp2'=e(N)*e(r2) scalar `lmhhp2p'=chi2tail(1, abs(`lmhhp2')) qui regress `E2' `LYh2' , `noconstant' scalar `lmhhp3'=e(N)*e(r2) scalar `lmhhp3p'= chi2tail(1, abs(`lmhhp3')) qui regress `absE' `SPXvar' , `noconstant' scalar `lmhgl'=e(mss)/((1-2/_pi)*`Sig2n') scalar `lmhglp'=chi2tail(2, abs(`lmhgl')) qui regress `U2' `Yh' , `noconstant' scalar `lmhcw1'= e(mss)/2 scalar `cwdf1' = e(df_m) scalar `lmhcw1p'=chi2tail(`cwdf1', abs(`lmhcw1')) qui regress `U2' `SPXvar' , `noconstant' scalar `lmhcw2'=e(mss)/2 scalar `cwdf2' =e(df_m) scalar `lmhcw2p'=chi2tail(`cwdf2', abs(`lmhcw2')) di di _dup(78) "{bf:{err:=}}" di as txt "{bf:{err:*** Panel Heteroscedasticity Tests}}" di _dup(78) "{bf:{err:=}}" di as txt _col(2) "{bf: Ho: Panel Homoscedasticity - Ha: Panel Heteroscedasticity}" di tempvar Time tempname lmharch lmharchp lmhbg lmhbgp qui gen `Time'=_n qui tsset `Time' qui cap drop `LE'* qui gen `LE'1=L1.`E2' qui regress `E2' `LE'* scalar `lmharch'=e(r2)*e(N) scalar `lmharchp'= chi2tail(1, abs(`lmharch')) qui regress `E2' L1.`E2' `SPXvar' , `noconstant' scalar `lmhbg'=e(r2)*e(N) scalar `lmhbgp'=chi2tail(1, abs(`lmhbg')) di as txt "- Engle LM ARCH Test AR(1): E2 = E2_1" _col(41) "=" as res %9.4f `lmharch' _col(54) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f `lmharchp' di _dup(78) "-" di as txt "- Hall-Pagan LM Test: E2 = Yh" _col(41) "=" as res %9.4f `lmhhp1' _col(54) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f `lmhhp1p' di as txt "- Hall-Pagan LM Test: E2 = Yh2" _col(41) "=" as res %9.4f `lmhhp2' _col(54) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f `lmhhp2p' di as txt "- Hall-Pagan LM Test: E2 = LYh2" _col(41) "=" as res %9.4f `lmhhp3' _col(54) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f `lmhhp3p' di _dup(78) "-" di as txt "- Harvey LM Test: LogE2 = X" _col(41) "=" as res %9.4f `lmhharv' _col(54) as txt "P-Value > Chi2(2)" _col(73) as res %5.4f `lmhharvp' di as txt "- Wald Test: LogE2 = X " _col(41) "=" as res %9.4f `lmhwald' _col(54) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f `lmhwaldp' di as txt "- Glejser LM Test: |E| = X" _col(41) "=" as res %9.4f `lmhgl' _col(54) as txt "P-Value > Chi2(2)" _col(73) as res %5.4f `lmhglp' di as txt "- Breusch-Godfrey Test: E = E_1 X" _col(41) "=" as res %9.4f `lmhbg' _col(54) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f `lmhbgp' di _dup(78) "-" di as txt "- Machado-Santos-Silva Test: Ev=Yh Yh2" _col(41) "=" as res %9.4f `lmhmss1' _col(54) as txt "P-Value > Chi2(" `mssdf1' ")" _col(73) as res %5.4f `lmhmss1p' di as txt "- Machado-Santos-Silva Test: Ev=X" _col(41) "=" as res %9.4f `lmhmss2' _col(54) as txt "P-Value > Chi2(" `mssdf2' ")" _col(73) as res %5.4f `lmhmss2p' ereturn scalar lmhmss2p= `lmhmss2p' ereturn scalar lmhmss2= `lmhmss2p' ereturn scalar lmhmss1p= `lmhmss1p' ereturn scalar lmhmss1= `lmhmss1' di _dup(78) "-" di as txt "- White Test - Koenker(R2): E2 = X" _col(41) "=" as res %9.4f `lmhw01' _col(54) as txt "P-Value > Chi2(" `dfw0' ")" _col(73) as res %5.4f `lmhw01p' di as txt "- White Test - B-P-G (SSR): E2 = X" _col(41) "=" as res %9.4f `lmhw02' _col(54) as txt "P-Value > Chi2(" `dfw0' ")" _col(73) as res %5.4f `lmhw02p' di _dup(78) "-" di as txt "- White Test - Koenker(R2): E2 = X X2" _col(41) "=" as res %9.4f `lmhw11' _col(54) as txt "P-Value > Chi2(" `dfw1' ")" _col(73) as res %5.4f `lmhw11p' di as txt "- White Test - B-P-G (SSR): E2 = X X2" _col(41) "=" as res %9.4f `lmhw12' _col(54) as txt "P-Value > Chi2(" `dfw1' ")" _col(73) as res %5.4f `lmhw12p' di _dup(78) "-" di as txt "- White Test - Koenker(R2): E2 = X X2 XX" _col(41) "=" as res %9.4f `lmhw21' _col(54) as txt "P-Value > Chi2(" `dfw2' ")" _col(73) as res %5.4f `lmhw21p' di as txt "- White Test - B-P-G (SSR): E2 = X X2 XX" _col(41) "=" as res %9.4f `lmhw22' _col(54) as txt "P-Value > Chi2(" `dfw2' ")" _col(73) as res %5.4f `lmhw22p' di _dup(78) "-" di as txt "- Cook-Weisberg LM Test: E2/S2n = Yh" _col(41) "=" as res %9.4f `lmhcw1' _col(54) as txt "P-Value > Chi2(" `cwdf1' ")" _col(73) as res %5.4f `lmhcw1p' di as txt "- Cook-Weisberg LM Test: E2/S2n = X" _col(41) "=" as res %9.4f `lmhcw2' _col(54) as txt "P-Value > Chi2(" `cwdf2' ")" _col(73) as res %5.4f `lmhcw2p' di _dup(78) "-" di as res "*** Single Variable Tests (E2/Sig2):" local nx : word count `SPXvar' tokenize `SPXvar' local i 1 while `i' <= `nx' { qui regress `U2' ``i'' ereturn scalar lmhcwx_`i'= e(mss)/2 ereturn scalar lmhcwxp_`i'= chi2tail(1 , abs(e(lmhcwx_`i'))) di as txt "- Cook-Weisberg LM Test: " "``i''" _col(44) "=" as res %9.4f e(lmhcwx_`i') _col(53) as txt " P-Value > Chi2(1)" _col(73) as res %5.4f e(lmhcwxp_`i') local i =`i'+1 } di _dup(78) "-" di as res "*** Single Variable Tests:" foreach i of local SPXvar { qui cap drop `ht'`i' tempvar `ht'`i' qui egen `ht'`i' = rank(`i') qui summ `ht'`i' scalar `mh' = r(mean) scalar `vh' = r(Var) qui summ `ht'`i' [aw=`E'^2] , meanonly scalar `h' = r(mean) ereturn scalar lmhq_`i'=(`N'^2/(2*(`N'-1)))*(`h'-`mh')^2/`vh' ereturn scalar lmhqp_`i'= chi2tail(1, abs(e(lmhq_`i'))) di as txt "- King LM Test: " "`i'" _col(44) "=" as res %9.4f e(lmhq_`i') _col(53) as txt " P-Value > Chi2(1)" _col(73) as res %5.4f e(lmhqp_`i') } di _dup(78) "-" tempvar Sig2 SigLR SigLRs SigLM SigLMs SigW SigWs E E2 EE1 En cN cT Obs Egh local idv `idv' qui tsset `Time' qui regress `yvar' `SPXvar' , `noconstant' qui predict `E' , res qui gen double `E2' = `E'^2 qui summ `E2' , meanonly local Sig2 = r(mean) qui gen double `SigLR' = . qui gen double `SigLM' = . qui gen double `SigW' = . local SigLRs = 0 local SigLMs = 0 local SigWs = 0 qui levelsof `idv' , local(levels) qui foreach l of local levels { summ `E2' if `idv' == `l', meanonly replace `SigLM'= (r(mean)/`Sig2'-1)^2 if `idv' == `l' replace `SigLR'= ln(r(mean))*r(N) if `idv' == `l' replace `SigW' = (`Sig2'/r(mean)-1)^2 if `idv' == `l' summ `SigLM' if `idv' == `l', meanonly local SigLMs =`SigLMs'+ r(mean) summ `SigLR' if `idv' == `l', meanonly local SigLRs = `SigLRs' + r(mean) summ `SigW' if `idv' == `l', meanonly local SigWs =`SigWs'+ r(mean) } local dflm= `NC'-1 local dflr= `NC'-1 local dfw = `NC' tempname lmhglr lmhglrp lmhglm lmhglmp lmhgw lmhgwp scalar `lmhglr'=`N'*ln(`Sig2')- `SigLRs' scalar `lmhglrp'= chi2tail(`dflr', abs(`lmhglr')) scalar `lmhglm'=`NT'/2*(`SigLMs') scalar `lmhglmp'= chi2tail(`dflm', abs(`lmhglm')) scalar `lmhgw'=`NT'/2*(`SigWs') scalar `lmhgwp'= chi2tail(`dfw', abs(`lmhgw')) di di _dup(78) "{bf:{err:=}}" di as txt "{bf:{err:* Panel Groupwise Heteroscedasticity Tests}}" di _dup(78) "{bf:{err:=}}" di as txt _col(2) "{bf: Ho: Panel Homoscedasticity - Ha: Panel Groupwise Heteroscedasticity}" di di as txt "- Lagrange Multiplier LM Test" _col(35) "=" as res %9.4f `lmhglm' as txt _col(50) "P-Value > Chi2(" `dflm' ")" _col(70) %5.4f as res `lmhglmp' di as txt "- Likelihood Ratio LR Test" _col(35) "=" as res %9.4f `lmhglr' _col(50) as txt "P-Value > Chi2(" `dflr' ")" _col(70) %5.4f as res `lmhglrp' di as txt "- Wald Test" as res _col(35) "=" as res %9.4f `lmhgw' _col(50) as txt "P-Value > Chi2(" `dfw' ")" _col(70) %5.4f as res `lmhgwp' di _dup(78) "-" ereturn scalar lmhglr=`lmhglr' ereturn scalar lmhglrp=`lmhglrp' ereturn scalar lmhglm=`lmhglm' ereturn scalar lmhglmp=`lmhglmp' ereturn scalar lmhgw=`lmhgw' ereturn scalar lmhgwp=`lmhgwp' ereturn scalar lmhw01=`lmhw01' ereturn scalar lmhw01p=`lmhw01p' ereturn scalar lmhw02=`lmhw02' ereturn scalar lmhw02p=`lmhw02p' ereturn scalar lmhw11=`lmhw11' ereturn scalar lmhw12=`lmhw12' ereturn scalar lmhw11p=`lmhw11p' ereturn scalar lmhw12p=`lmhw12p' ereturn scalar lmhw21=`lmhw21' ereturn scalar lmhw22=`lmhw22' ereturn scalar lmhw21p=`lmhw21p' ereturn scalar lmhw22p=`lmhw22p' ereturn scalar lmhharv=`lmhharv' ereturn scalar lmhharvp=`lmhharvp' ereturn scalar lmhwald=`lmhwald' ereturn scalar lmhwaldp=`lmhwaldp' ereturn scalar lmhhp1=`lmhhp1' ereturn scalar lmhhp1p=`lmhhp1p' ereturn scalar lmhhp2=`lmhhp2' ereturn scalar lmhhp2p=`lmhhp2p' ereturn scalar lmhhp3=`lmhhp3' ereturn scalar lmhhp3p=`lmhhp3p' ereturn scalar lmhgl=`lmhgl' ereturn scalar lmhglp=`lmhglp' ereturn scalar lmhcw1=`lmhcw1' ereturn scalar lmhcw1p=`lmhcw1p' ereturn scalar lmhcw2=`lmhcw2' ereturn scalar lmhcw2p=`lmhcw2p' ereturn scalar lmharch=`lmharch' ereturn scalar lmharchp=`lmharchp' ereturn scalar lmhbg=`lmhbg' ereturn scalar lmhbgp=`lmhbgp' } if "`lmnorm'"!= "" { tempvar Yh E E1 E2 E3 E4 Es U2 DE LDE LDF1 Yt U Hat tempname Hat corr1 corr3 corr4 mpc2 mpc3 mpc4 s uinv q1 uinv2 q2 ECov ECov2 Eb Sk Ku tempname M2 M3 M4 K2 K3 K4 Ss Kk GK sksd kusd N1 N2 EN S2N SN mean sd small A2 B0 B1 tempname B2 B3 LA Z Rn Lower Upper wsq2 ve lve Skn gn an cn kn vz Ku1 Kun n1 n2 n3 eb2 tempname R2W vb2 svb2 k1 a devsq m2 sdev m3 m4 sqrtb1 b2 g1 g2 stm3b2 S1 S2 S3 S4 tempname b2minus3 sm sms y k2 wk delta alpha yalpha pc1 pc2 pc3 pc4 pcb1 pcb2 sqb1p b2p qui tsset `Time' qui gen `Yh'=`Yh_ML' qui gen `E' =`Ue_ML' qui gen `E2'=`E'*`E' matrix `Hat'=vecdiag(`X'*invsym(`X''*`X')*`X'')' qui svmat `Hat' , name(`Hat') rename `Hat'1 `Hat' qui regress `E2' `Hat' scalar `R2W'=e(r2) qui summ `E' , det scalar `Eb'=r(mean) scalar `Sk'=r(skewness) scalar `Ku'=r(kurtosis) qui forvalue i = 1/4 { qui gen `E'`i'=(`E'-`Eb')^`i' qui summ `E'`i' scalar `S`i''=r(mean) scalar `pc`i''=r(sum) } mkmat `E'1 `E'2 `E'3 `E'4 in 1/`N' , matrix(`ECov') scalar `M2'=`S2'-`S1'^2 scalar `M3'=`S3'-3*`S1'*`S2'+`S1'^2 scalar `M4'=`S4'-4*`S1'*`S3'+6*`S1'^2*`S2'-3*`S1'^4 scalar `K2'=`N'*`M2'/(`N'-1) scalar `K3'=`N'^2*`M3'/((`N'-1)*(`N'-2)) scalar `K4'=`N'^2*((`N'+1)*`M4'-3*(`N'-1)*`M2'^2)/((`N'-1)*(`N'-2)*(`N'-3)) scalar `Ss'=`K3'/(`K2'^1.5) scalar `Kk'=`K4'/(`K2'^2) scalar `GK'= ((`Sk'^2/6)+((`Ku'-3)^2/24)) ereturn scalar lmnw=`N'*(`R2W'+`GK') ereturn scalar lmnwp= chi2tail(2, abs(e(lmnw))) ereturn scalar lmnjb=`N'*((`Sk'^2/6)+((`Ku'-3)^2/24)) ereturn scalar lmnjbp= chi2tail(2, abs(e(lmnjb))) scalar `sksd'=sqrt(6*`N'*(`N'-1)/((`N'-2)*(`N'+1)*(`N'+3))) scalar `kusd'=sqrt(24*`N'*(`N'-1)^2/((`N'-3)*(`N'-2)*(`N'+3)*(`N'+5))) qui gen `DE'=1 if `E'>0 qui replace `DE'=0 if `E' <= 0 qui count if `DE'>0 scalar `N1'=r(N) scalar `N2'=`N'-r(N) scalar `EN'=(2*`N1'*`N2')/(`N1'+`N2')+1 scalar `S2N'=(2*`N1'*`N2'*(2*`N1'*`N2'-`N1'-`N2'))/((`N1'+`N2')^2*(`N1'+`N2'-1)) scalar `SN'=sqrt((2*`N1'*`N2'*(2*`N1'*`N2'-`N1'-`N2'))/((`N1'+`N2')^2*(`N1'+`N2'-1))) qui gen `LDE'= `DE'[_n-1] qui replace `LDE'=0 if `DE'==1 in 1 qui gen `LDF1'= 1 if `DE' != `LDE' qui replace `LDF1'= 1 if `DE' == `LDE' in 1 qui replace `LDF1'= 0 if `LDF1' == . qui count if `LDF1'>0 scalar `Rn'=r(N) ereturn scalar lmng=(`Rn'-`EN')/`SN' ereturn scalar lmngp= chi2tail(2, abs(e(lmng))) scalar `Lower'=`EN'-1.96*`SN' scalar `Upper'=`EN'+1.96*`SN' qui summ `E' scalar `mean'=r(mean) scalar `sd'=r(sd) scalar `small'= 1e-20 qui gen `Es' =`E' qui sort `Es' qui replace `Es'=normal((`Es'-`mean')/`sd') qui gen `Yt'=`Es'*(1-`Es'[`N'-_n+1]) qui replace `Yt'=`small' if `Yt' < =0 qui replace `Yt'=sum((2*_n-1)*ln(`Yt')) scalar `A2'=-`N'-`Yt'[`N']/`N' scalar `A2'=`A2'*(1+(0.75+2.25/`N')/`N') scalar `B0'=2.25247+0.000317*exp(29.5/`N') scalar `B1'=2.16872+0.00243*exp(27.7/`N') scalar `B2'=0.19135+0.00255*exp(28.3/`N') scalar `B3'=0.110978+0.00001624*exp(39.04/`N')+0.00476*exp(21.37/`N') scalar `LA'=ln(`A2') ereturn scalar lmnad=(`A2') scalar `Z'=abs(`B0'+`LA'*(`B1'+`LA'*(`B2'+`LA'*`B3'))) ereturn scalar lmnadp= normal(abs(-`Z')) scalar `wsq2'=-1+sqrt(2*((3*(`N'^2+27*`N'-70)/((`N'-2)*(`N'+5))*((`N'+1)/(`N'+7))*((`N'+3)/(`N'+9)))-1)) scalar `ve'=`Sk'*sqrt((`N'+1)*(`N'+3)/(6*(`N'-2)))/sqrt(2/(`wsq2'-1)) scalar `lve'=ln(`ve'+(`ve'^2+1)^0.5) scalar `Skn'=`lve'/sqrt(ln(sqrt(`wsq2'))) scalar `gn'=((`N'+5)/(`N'-3))*((`N'+7)/(`N'+1))/(6*(`N'^2+15*`N'-4)) scalar `an'=(`N'-2)*(`N'^2+27*`N'-70)*`gn' scalar `cn'=(`N'-7)*(`N'^2+2*`N'-5)*`gn' scalar `kn'=(`N'*`N'^2+37*`N'^2+11*`N'-313)*`gn'/2 scalar `vz'= `cn'*`Sk'^2 +`an' scalar `Ku1'=(`Ku'-1-`Sk'^2)*`kn'*2 scalar `Kun'=(((`Ku1'/(2*`vz'))^(1/3))-1+1/(9*`vz'))*sqrt(9*`vz') ereturn scalar lmndh =`Skn'^2 + `Kun'^2 ereturn scalar lmndhp= chi2tail(2, abs(e(lmndh))) scalar `n1'=sqrt(`N'*(`N'-1))/(`N'-2) scalar `n2'=3*(`N'-1)/(`N'+1) scalar `n3'=(`N'^2-1)/((`N'-2)*(`N'-3)) scalar `eb2'=3*(`N'-1)/(`N'+1) scalar `vb2'=24*`N'*(`N'-2)*(`N'-3)/(((`N'+1)^2)*(`N'+3)*(`N'+5)) scalar `svb2'=sqrt(`vb2') scalar `k1'=6*(`N'*`N'-5*`N'+2)/((`N'+7)*(`N'+9))*sqrt(6*(`N'+3)*(`N'+5)/(`N'*(`N'-2)*(`N'-3))) scalar `a'=6+(8/`k1')*(2/`k1'+sqrt(1+4/(`k1'^2))) scalar `devsq'=`pc1'*`pc1'/`N' scalar `m2'=(`pc2'-`devsq')/`N' scalar `sdev'=sqrt(`m2') scalar `m3'=`pc3'/`N' scalar `m4'=`pc4'/`N' scalar `sqrtb1'=`m3'/(`m2'*`sdev') scalar `b2'=`m4'/`m2'^2 scalar `g1'=`n1'*`sqrtb1' scalar `g2'=(`b2'-`n2')*`n3' scalar `stm3b2'=(`b2'-`eb2')/`svb2' ereturn scalar lmnkz=(1-2/(9*`a')-((1-2/`a')/(1+`stm3b2'*sqrt(2/(`a'-4))))^(1/3))/sqrt(2/(9*`a')) ereturn scalar lmnkzp=2*(1-normal(abs(e(lmnkz)))) scalar `b2minus3'=`b2'-3 matrix `ECov2'=`ECov''*`ECov' scalar `sm'=`ECov2'[1,1] scalar `sms'=1/sqrt(`sm') matrix `corr1'=`sms'*`ECov2'*`sms' matrix `corr3'=`corr1'[1,1]^3 matrix `corr4'=`corr1'[1,1]^4 scalar `y'=`sqrtb1'*sqrt((`N'+1)*(`N'+3)/(6*(`N'-2))) scalar `k2'=3*(`N'^2+27*`N'-70)*(`N'+1)*(`N'+3)/((`N'-2)*(`N'+5)*(`N'+7)*(`N'+9)) scalar `wk'=sqrt(sqrt(2*(`k2'-1))-1) scalar `delta'=1/sqrt(ln(`wk')) scalar `alpha'=sqrt(2/(`wk'*`wk'-1)) matrix `yalpha'=`y'/`alpha' scalar `yalpha'=`yalpha'[1,1] ereturn scalar lmnsz=`delta'*ln(`yalpha'+sqrt(1+`yalpha'^2)) ereturn scalar lmnszp= 2*(1-normal(abs(e(lmnsz)))) ereturn scalar lmndp=e(lmnsz)^2+e(lmnkz)^2 ereturn scalar lmndpp= chi2tail(2, abs(e(lmndp))) matrix `uinv'=invsym(`corr3') matrix `q1'=e(lmnsz)'*`uinv'*e(lmnsz) ereturn scalar lmnsms=`q1'[1,1] ereturn scalar lmnsmsp= chi2tail(1, abs(e(lmnsms))) matrix `uinv2'=invsym(`corr4') matrix `q2'=e(lmnkz)'*`uinv2'*e(lmnkz) ereturn scalar lmnsmk=`q2'[1,1] ereturn scalar lmnsmkp= chi2tail(1, abs(e(lmnsmk))) matrix `mpc2'=(`pc2'-(`pc1'^2/`N'))/`N' matrix `mpc3'=(`pc3'-(3/`N'*`pc1'*`pc2')+(2/(`N'^2)*(`pc1'^3)))/`N' matrix `mpc4'=(`pc4'-(4/`N'*`pc1'*`pc3')+(6/(`N'^2)*(`pc2'*(`pc1'^2)))-(3/(`N'^3)*(`pc1'^4)))/`N' scalar `pcb1'=`mpc3'[1,1]/`mpc2'[1,1]^1.5 scalar `pcb2'=`mpc4'[1,1]/`mpc2'[1,1]^2 scalar `sqb1p'=`pcb1'^2 scalar `b2p'=`pcb2' ereturn scalar lmnsvs=`sqb1p'*`N'/6 ereturn scalar lmnsvsp= chi2tail(1, abs(e(lmnsvs))) ereturn scalar lmnsvk=(`b2p'-3)*sqrt(`N'/24) ereturn scalar lmnsvkp= 2*(1-normal(abs(e(lmnsvk)))) qui sort `Time' di di _dup(78) "{bf:{err:=}}" di as txt "{bf:{err:* Panel Non Normality Tests}}" di _dup(78) "{bf:{err:=}}" di as txt "{bf: Ho: Normality - Ha: Non Normality}" di _dup(78) "-" di "{bf:*** Non Normality Tests:} di as txt "- Jarque-Bera LM Test" _col(40) "=" as res %9.4f e(lmnjb) _col(55) as txt "P-Value > Chi2(2) " _col(73) as res %5.4f e(lmnjbp) di as txt "- White IM Test" _col(40) "=" as res %9.4f e(lmnw) _col(55) as txt "P-Value > Chi2(2) " _col(73) as res %5.4f e(lmnwp) di as txt "- Doornik-Hansen LM Test" _col(40) "=" as res %9.4f e(lmndh) _col(55) as txt "P-Value > Chi2(2) " _col(73) as res %5.4f e(lmndhp) di as txt "- Geary LM Test" _col(40) "=" as res %9.4f e(lmng) _col(55) as txt "P-Value > Chi2(2) " _col(73) as res %5.4f e(lmngp) di as txt "- Anderson-Darling Z Test" _col(40) "=" as res %9.4f e(lmnad) _col(55) as txt "P > Z(" %6.3f `Z' ")" _col(73) as res %5.4f e(lmnadp) di as txt "- D'Agostino-Pearson LM Test " _col(40) "=" as res %9.4f e(lmndp) _col(55) as txt "P-Value > Chi2(2)" _col(73) as res %5.4f e(lmndpp) di _dup(78) "-" di "{bf:*** Skewness Tests:} di as txt "- Srivastava LM Skewness Test" _col(40) "=" as res %9.4f e(lmnsvs) _col(55) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f e(lmnsvsp) di as txt "- Small LM Skewness Test" _col(40) "=" as res %9.4f e(lmnsms) _col(55) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f e(lmnsmsp) di as txt "- Skewness Z Test" _col(40) "=" as res %9.4f e(lmnsz) _col(55) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f e(lmnszp) di _dup(78) "-" di "{bf:*** Kurtosis Tests:} di as txt "- Srivastava Z Kurtosis Test" _col(40) "=" as res %9.4f e(lmnsvk) _col(55) as txt "P-Value > Z(0,1)" _col(73) as res %5.4f e(lmnsvkp) di as txt "- Small LM Kurtosis Test" _col(40) "=" as res %9.4f e(lmnsmk) _col(55) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f e(lmnsmkp) di as txt "- Kurtosis Z Test" _col(40) "=" as res %9.4f e(lmnkz) _col(55) as txt "P-Value > Chi2(1)" _col(73) as res %5.4f e(lmnkzp) di _dup(78) "-" di as txt _col(5) "Skewness Coefficient =" _col(28) as res %7.4f `Sk' as txt " " " - Standard Deviation = " _col(48) as res %7.4f `sksd' di as txt _col(5) "Kurtosis Coefficient =" _col(28) as res %7.4f `Ku' as txt " " " - Standard Deviation = " _col(48) as res %7.4f `kusd' di _dup(78) "-" di as txt _col(5) "Runs Test:" as res " " "(" `Rn' ")" " " as txt "Runs - " as res " " "(" `N1' ")" " " as txt "Positives -" " " as res "(" `N2' ")" " " as txt "Negatives" di as txt _col(5) "Standard Deviation Runs Sig(k) = " as res %7.4f `SN' " , " as txt "Mean Runs E(k) = " as res %7.4f `EN' di as txt _col(5) "95% Conf. Interval [E(k)+/- 1.96* Sig(k)] = (" as res %7.4f `Lower' " , " %7.4f `Upper' " )" di _dup(78) "-" } if "`tolog'"!="" { qui foreach var of local vlistlog { qui replace `var'= `xyind`var'' } } if inlist("`mfx'", "lin", "log") { tempname mfxb mfxe mfxlin mfxlog XMB XYMB YMB YMB1 qui mean `SPXvar' matrix `XMB'=e(b)' qui summ `yvar' scalar `YMB1'=r(mean) matrix `YMB'=J(rowsof(`XMB'),1,`YMB1') mata: X = st_matrix("`XMB'") mata: Y = st_matrix("`YMB'") if inlist("`mfx'", "lin") { mata: `XYMB'=X:/Y mata: `XYMB'=st_matrix("`XYMB'",`XYMB') matrix `mfxb' =`Bx' matrix `mfxe'=vecdiag(`Bx'*`XYMB'')' matrix `mfxlin' =`mfxb',`mfxe',`XMB' matrix rownames `mfxlin' = `SPXvar' matrix colnames `mfxlin' = Marginal_Effect(B) Elasticity(Es) Mean matlist `mfxlin' , title({bf:* Marginal Effect - Elasticity {bf:(Model= {err:`model'})}: {err:Linear} *}) twidth(10) border(all) lines(columns) rowtitle(Variable) format(%18.4f) ereturn matrix mfxlin=`mfxlin' } if inlist("`mfx'", "log") { mata: `XYMB'=Y:/X mata: `XYMB'=st_matrix("`XYMB'",`XYMB') matrix `mfxe'=`Bx' matrix `mfxb'=vecdiag(`Bx'*`XYMB'')' matrix `mfxlog' =`mfxe',`mfxb',`XMB' matrix rownames `mfxlog' = `SPXvar' matrix colnames `mfxlog' = Elasticity(Es) Marginal_Effect(B) Mean matlist `mfxlog' , title({bf:* Elasticity - Marginal Effect {bf:(Model= {err:`model'})}: {err:Log-Log} *}) twidth(10) border(all) lines(columns) rowtitle(Variable) format(%18.4f) ereturn matrix mfxlog=`mfxlog' } di as txt " Mean of Dependent Variable =" as res _col(30) %12.4f `YMB1' } qui cap matrix drop _all qui cap mata: mata drop * if "`zero'"!="" { qui foreach var of local varlist { qui replace `var'= `zeromiss'`var' } } qui sort `Time' end program define yxregeq version 10.0 syntax varlist gettoken yvar xvar : varlist local LEN=length("`yvar'") local LEN=`LEN'+3 di "{p 2 `LEN' 5}" as res "{bf:`yvar'}" as txt " = " " local kx : word count `xvar' local i=1 while `i'<=`kx' { local X : word `i' of `xvar' if `i'<`kx' { di " " as res " {bf:`X'}" _c di as txt " + " _c } if `i'==`kx' { di " " as res "{bf:`X'}" } local i=`i'+1 } di "{p_end}" di as txt "{hline 78}" end