Table 4.
ARDL model results.
| Selected Model |
ARDL (3, 0, 1, 1, 1, 0, 0) |
ARDL (1, 0, 0, 0, 0, 1, 1) |
ARDL (1, 0, 0, 0, 0, 1, 1) |
||
|---|---|---|---|---|---|
| Dependent Variable | ln CO2 | ln CH4 | ln N2O | ||
| Long run coefficients | |||||
| Scale effect ln RGDP |
1.248∗ |
0.4224∗ |
0.543∗ |
||
| Composition effect ln Ind-ratio (ln Ind-ratio)2 |
3.985∗ 0.985∗ |
-0.533∗∗ -0.118 |
-0.488∗∗ -0.061 |
||
| Technique effect6 ln GNDI (ln GNDI)2 |
-0.167 -0.039∗ |
1.201∗ -0.077∗ |
0.803∗ -0.057∗ |
||
| ln FEC | 1.399∗ | 0.259∗∗ | 0.236∗∗ | ||
| Short run coefficients | |||||
| Scale effect Δ(ln RGDP) |
0.965∗ |
0.198∗ |
0.372∗ |
||
| Composition effect Δ (ln Ind-ratio) Δ (ln Ind-ratio)2 |
1.125 -0.183 |
-0.249∗∗ -0.055∗∗∗ |
-0.308∗∗∗ -0.042 |
||
| Technique effect Δ (ln GNDI) Δ (ln GNDI)2 |
-1.085∗ -0.031∗ |
0.562∗ -0.007∗∗∗ |
0.551∗ -0.010∗∗ |
||
| Δ (ln FEC) | 0.727∗ | 0.036 | 0.056 | ||
| Error correction term | -0.774∗ | -0.468∗ | -0.686∗ | ||
| Model Diagnostic tests | |||||
| R2 [adj-R2] |
0.992 [0.989] |
0.991 [0.989] |
0.985 [0.982] |
||
| Breusch-Godfrey Serial Correlation LM Test [P-Value] | 0.130 [0.878] |
2.335 [0.114] |
0.267 [0.609] |
||
| Heteroscedasticity Test: ARCH [P-Value] | 1.036 [0.315] |
0.274 [0.603] |
0.089 [0.766] |
||
| Ramsey RESET Test [P-Value] |
0.056 [0.814] |
0.294 [0.592] |
1.416 [0.243] |
||
| Normality: JB test [p-value] |
0.056 [0.972] |
3.058 [0.217] |
0.111 [0.946] |
||
Note: ∗, ∗∗ and ∗∗∗ denote significance at 1 %, 5 % and 10 % levels, respectively. The optimal lag structure for each model is automatically selected based on Akaike Information Criteria (AIC).