现在网上很少能找到用蒙特卡洛模拟生成随机变量，并且带入时间序列的代码，大部分还都是蒙特卡洛和线性回归。
这是我自己参考了以下paper写得代码，供大家参考一下。 [size=12.000000pt]MONTE CARLO EXPERIMENTS USING STATA: A PRIMER WITHEXAMPLES
[size=12.000000pt]

这是需要模拟的AR(1)：[LaTex]y_{t}=b_{0}+b_{1}y_{t-1}+u_{t}[/LaTex]


[LaTex]Y\sim N(\mu ,\sigma ^{2})[/LaTex]
[LaTex]u_{t}\sim N(0,1)[/LaTex]
R=1000, Sample Size 400
[size=12.000000pt]

global nobs=400

global nmc=1000

set seed 10000

set obs $nobs

gen time= _n

tsset time

scalar constant = 1

scalar delta = 0.9

scalar sigma = 1

gen u=0

gen y=0

program regLDV2, rclass

tempname sim

postfile ‘sim’ b1se1using results, replace

quietly{

forvalues i = 1/$nmc{

replace u = rnormal(0,sigma)

replace y=rnormal(10,100/19)

replace y = constant + delta *L.y + u in 2/$nobs

reg y L.y

scalar b1 =_b[L.y]

scalar se1=_se[L.y]

post ‘sim’ (b1) (se1)

}

}

postclose ‘sim’

end

regLDV2

use results, clear

summarize

sktest b1 se1

clear