g t = year - 2005
g alpha = nrK/(nwL+nrK)
replace nwL = nY * (1-alpha)
replace nrK = nY * alpha
foreach var in nY nK nL alpha {
  ameans `var'
  scalar `var'_bar = r(mean_g)
}
  ameans t
  scalar t_bar = r(mean)
scalar list
g nY_norm = nY/nY_bar
g nL_norm = nL/nL_bar
g nK_norm = nK/nK_bar
g t_norm = t/t_bar
g t_dif = t-t_bar

g lLs = log(1-alpha)
g lKs = log(alpha)
g lnY_norm = log(nY_norm)
nlsur (lnY_norm = log({xi}) + 1/(({e}-1)/{e})*log(((1-{a})*(nL_norm*exp(t_bar*{gammaL}/{lammaL}*(t_norm^{lammaL}-1)))^(({e}-1)/{e})) + ({a}*(nK_norm*exp(t_bar*{gammaK}/{lammaK}*(t_norm^{lammaK}-1)))^(({e}-1)/{e})))    ) ///
     (lLs = log(1-{a})  +(({e}-1)/{e})*log({xi}) - (({e}-1)/{e})*log(nY_norm/nL_norm) + (({e}-1)/{e})*(t_bar*{gammaL}/{lammaL}*(t_norm^{lammaL}-1))) ///
     (lKs = log({a})  +(({e}-1)/{e})*log({xi}) - (({e}-1)/{e})*log(nY_norm/nK_norm) + (({e}-1)/{e})*(t_bar*{gammaK}/{lammaK}*(t_norm^{lammaK}-1))) ///
if province==1, initial(e 0.2 xi 1 lammaL 1 lammaK 1  a 0.5)
estimates store all
g delta = (nrK/nK) / (nwL/nL)
scalar i = 0
matrix ela = (0.74, 0.505)
scalar sc_a = 0.39
forvalue i = 1(1)2 {
g nBA`i' = nL/nK * ((1-sc_a)/sc_a * alpha/(1-alpha))^(ela[1,`i']/(ela[1,`i'] - 1))
g nD`i' = 1/delta * sc_a/(1 - sc_a) * (1-1/ela[1,`i']) * ( (nBA`i'+nBA`i'[_n-1])/2 )^(-1/ela[1,`i']) * ((nL/nK+ nL[_n-1]/nK[_n-1])/2 )^(1/ela[1,`i']) * (nBA`i' - nBA`i'[_n-1])
}
list nD*
这是戴天仕(2010)计算技术进步方向的代码，但是截面数据，小弟200多个地区的数据，请问该怎么写循环处理