I have done the first step OLS and generated the residual seriesRESID01

my Equation01: logdax c logdax(-1 to -3) normal OLS estimation

At the new window of RESID01 , i do the correl. test with 1 order and 2 order difference , it is show me that there are also very strongly correlation at both order(-0.50) another (-0.671), what does it means? Lag operator of first oder and second order are all significant strong!

Under the equation 01, at residual test, it showed me that there are significant corr. at squared residual(all significant less than 0.25)

What should i do at third step?

Could i estimate with a new equation

Logdax c logdax(-1 to -3) directly with method ARCH and let GARCH(1),or ARCH(1)

could someone explain it?