Hello. I am trying to do error correction model with rolling regression.
here. I want to change the first step as yt= bo + xt (setting the coefficient as always 1, with rolling regression way t=1~40, 2~41, ..... 128~167)
Can you please give me some advices on this issue?
Thank you in advance!

-------------------------------------------------------------------------------------------------------------------------------------------------------
1. first step
forval i=1/167{

reg y x if t<`=39+`i'' & t>`=`i'-1'

predict uhat`i' if t<`=39+`i'' & t>`=`i'-1', res

}

​

2. second step

forval i=1/167{

reg D.y L.uhat`i' D.x if t<`=39+`i'' & t>`=`i'-1'

ereturn list

mat coef`i'=e(b)

}

​
​

mat comb = (coef1 \ coef2 \ coef3 \ coef4 \ coef5 \ coef6 \ coef7 \ coef8 \ coef9 \ coef10 \ coef11 \ coef12 \ coef13 \ coef14 \ coef15 \ coef16 \ coef17 \ coef18 \ coef19 \ coef20 \ coef21 \ coef22 \ coef23 \ coef24 \ coef25 \ coef26 \ coef27 \ coef28 \ coef29 \ coef30 \ coef31 \ coef32 \ coef33 \ coef34 \ coef35 \ coef36 \ coef37 \ coef38 \ coef39 \ coef40 \ coef41 \ coef42 \ coef43 \ coef44 \ coef45 \ coef46 \ coef47 \ coef48 \ coef49 \ coef50 \ coef51 \ coef52 \ coef53 \ coef54 \ coef55 \ coef56 \ coef57 \ coef58 \ coef59 \ coef60 \ coef61 \ coef62 \ coef63 \ coef64 \ coef65 \ coef66 \ coef67 \ coef68 \ coef69 \ coef70 \ coef71 \ coef72 \ coef73 \ coef74 \ coef75 \ coef76 \ coef77 \ coef78 \ coef79 \ coef80 \ coef81 \ coef82 \ coef83 \ coef84 \ coef85 \ coef86 \ coef87 \ coef88 \ coef89 \ coef90 \ coef91 \ coef92 \ coef93 \ coef94 \ coef95 \ coef96 \ coef97 \ coef98 \ coef99 \ coef100 \ coef101 \ coef102 \ coef103 \ coef104 \ coef105 \ coef106 \ coef107 \ coef108 \ coef109 \ coef110 \ coef111 \ coef112 \ coef113 \ coef114 \ coef115 \ coef116 \ coef117 \ coef118 \ coef119 \ coef120 \ coef121 \ coef122 \ coef123 \ coef124 \ coef125 \ coef126 \ coef127 \ coef128 \ coef129)

mat list comb

svmat comb

​

mat list comb

svmat comb

-------------------------------------------------------------------------------------------------------------------------------------------------------