Explanation

To test for ARCH(1) effects in the residuals:

Correct procedure:

1. Estimate the squared residuals from the AR(2) model
2. Regress these squared residuals on:
   • A constant term
   • One lag of the squared residuals (for ARCH(1) testing)
3. Test whether the slope coefficient is statistically different from 0

Why this works:

• ARCH(1) means the conditional variance depends on the previous period's squared error
• If the slope coefficient is significantly different from zero, it indicates ARCH effects
• Testing against 1 would be incorrect - we're testing for the presence of ARCH, not specific parameter values

This is the standard Engle's ARCH test procedure for detecting autoregressive conditional heteroskedasticity.