Bootstrapping dynamic threshold panel model error

[Jane Ho]

Hi,

I am trying to estimate the threshold level of human capital on productivity growth, using the dynamic panel threshold method introduced by Dang, Kim and Shin (2012), Kremer et al (2013) and Chao et al (2017). Here are some initial preliminaries:

. local vary "y" //productivity growth

. local vardx "dm df dh de dp" //explanatory variables

. local vartd "td2 td3 td4 td5 td6 td7 td8" //time dummies

. local nb = 1000 //number of bootstrapping


However, when I go to the bootstrapping test, the following error appears:


. dis "Start bootstrapping"
Start bootstrapping

. forvalues i= 1(1)`nb'{
2. dis `i'
3. mata: upsamp_regime("no","time","nobs","subno","`vary'", "`vardx'","`vartd'","dehat","b0")
4. qui gen lyb1 = l.yb*i1
5. qui gen lyb2 = l.yb*i2
6. qui xtabond2 yb lyb1 lyb2 m1 f1 h1 e1 p1 m2 f2 h2 e2 p2 td? cons, gmm(lyb1 lyb2) iv( m1 f1 h1 e1 p1 m2 f2 h2 e2 p2) iv(td? cons) nolevel robust twostep
7. matrix bb= e(b)'
8. matrix vb = e(V)
9. matrix W = (H*bb)'*inv(H*vb*H')*(H*bb)
10. scalar WaldStatsb = W[1,1]
11. matrix WBOOT[`i',1]= WaldStatsb
12. if WaldStats0 < WaldStatsb{
13. scalar pv = pv+1
14. }
15. drop yb lyb1 lyb2
16. }
1
upsamp_regime(): 3301 subscript invalid
<istmt>: - function returned error
r(3301);

Can anyone give me some pieces of advice how to solve this problem, please?

Thanks much!!!

Jane

[fai ot]

Hi,

I am working on dynamic panel with threshold effect method of Seo and Shin(2014) , do you have any idea about stata command of this method?

Thanks
Fairouz.