Table 3.
Autoregressive conditional heteroskedasticity test results
| Model (lag length years) | Akaike information criterion | F statistic (P value) | Observation × R2 (P value) |
|---|---|---|---|
| ARCH (4) | 6.66 | 1.10 (0.392) | 4.54 (0.338) |
| ARCH (3) | 6.69 | 2.21 (0.125) | 5.88 (0.117) |
| ARCH (2) | 8.01 | 0.51 (0.607) | 1.13 (0.569) |
| ARCH (1) | 7.92 | 0.29 (0.597) | 0.31 (0.577) |
ARCH = Autoregressive conditional heteroskedasticity.
The ARCH test was used to test the heteroskedasticity of the residual in the selected regression model. The time lag of the ARCH test was set to 1, 2, 3, and 4 years. The significance levels of the F statistic, and observation × R2 values were set to 0.05.