2

I am new to econometrics and I am building my first econometric model. I ran the LM test on a univariate time series data of 12000 observations and got the following stats:

Chi-squared = 458.5313, df = 12, p-value < 2.2e-16

Can I reject the null? Is there presence of ARCH effects?

Andre Silva
  • 3,070
  • 5
  • 28
  • 55
Piscean
  • 21
  • 1
  • 3

1 Answers1

2

Can I reject the null?

With a p-value below .Machine$double.eps you would reject the null hypothesis for the Lagrange multiplier test at every reasonable significance level.

If your assumptions hold, the null hypothesis doesn't seem to be tenable

Is there presence of ARCH effects?

That might be the obvious conclusion, but rejection may occur for other reasons than the null being false in exactly the way we write the alternative. It's presumably possible, for example, for a non-ARCH process to produce a significant result here.

Glen_b
  • 257,508
  • 32
  • 553
  • 939