Hello, we have monthly returns from 3 sectoral indexes from a country (r_bvl_ind r_bvl_min r_bvl_ser) and the monthly returns from the S&P500 (r_sp500), we want to apply a DCC-GARCH model in order to analyze the volatility transmissions from the S&P 500 to these sectors. Could someone help us with the stata scripts?
gen t = year(Date)
gen monthnum = month(Date)
gen mdate = ym(t, monthnum)
format mdate %tm
sort mdate
tsset mdate, monthly
* ---Period 1: Jan 1999 – Apr 2015 ---
preserve
keep if mdate <= tm(2015m4)
arch r_bvl_ind, arch(1) garch(1) technique(bhhh)
est store ind_1
arch r_bvl_min, arch(1) garch(1) technique(bfgs)
est store min_1
arch r_bvl_fin, arch(1) garch(1) technique(bhhh)
est store fin_1
restore
* --- Period 2: May 2015 – Aug 2025 ---
preserve
keep if mdate > tm(2015m4)
arch r_bvl_ind, arch(1) garch(1) technique(bfgs)
est store ind_2
arch r_bvl_min, arch(1) garch(1) technique(bfgs)
est store min_2
arch r_bvl_fin, arch(1) garch(1) technique(bhhh)
est store fin_2
restore
We tried mgarch dcc ( r_bvl_ind r_bvl_fin r_bvl_fin = r_sp500), arch (1 2)
But an error appears "r_bvl_fin collinear with other dependent variables"
How should we proceed? Thanks in advance.
gen t = year(Date)
gen monthnum = month(Date)
gen mdate = ym(t, monthnum)
format mdate %tm
sort mdate
tsset mdate, monthly
* ---Period 1: Jan 1999 – Apr 2015 ---
preserve
keep if mdate <= tm(2015m4)
arch r_bvl_ind, arch(1) garch(1) technique(bhhh)
est store ind_1
arch r_bvl_min, arch(1) garch(1) technique(bfgs)
est store min_1
arch r_bvl_fin, arch(1) garch(1) technique(bhhh)
est store fin_1
restore
* --- Period 2: May 2015 – Aug 2025 ---
preserve
keep if mdate > tm(2015m4)
arch r_bvl_ind, arch(1) garch(1) technique(bfgs)
est store ind_2
arch r_bvl_min, arch(1) garch(1) technique(bfgs)
est store min_2
arch r_bvl_fin, arch(1) garch(1) technique(bhhh)
est store fin_2
restore
We tried mgarch dcc ( r_bvl_ind r_bvl_fin r_bvl_fin = r_sp500), arch (1 2)
But an error appears "r_bvl_fin collinear with other dependent variables"
How should we proceed? Thanks in advance.

Comment