ARCH and Trend Models

Hey AF

  1. If and ARCH test shows a statistically significant slope on the lagged squared results, thus that ARCH is present for an ARCH(1) model, does this have implications for my AR(1) trend model?

  2. Maybe I got it mixed up and should only focus on testing whether the residual terms exhibit no autocorrelation for the AR-model or does ARCH tell us that the standard errors in an AR model will be too small due to conditional heteroskedasticity?

Kind regards,
Christoffer