Table 5.

Non-linear panel ARDL analysis.

Variables	Dependent variable: COVID-19 cases
ECT	−0.123***
	(0.0274)
$\Delta \mathit{P}{\mathit{M}}_{2.5}\mathit{pollutio}{\mathit{n}}_{\mathit{t}}^{+}$	0.112**
	(0.0554)
$\Delta \mathit{P}{\mathit{M}}_{2.5}\mathit{pollutio}{\mathit{n}}_{\mathit{t}}^{-}$	−0.0348
	(0.0369)
$\mathit{P}{\mathit{M}}_{2.5}\mathit{pollutio}{\mathit{n}}_{\mathit{t}}^{+}$	0.439***
	(0.144)
$\mathit{P}{\mathit{M}}_{2.5}\mathit{pollutio}{\mathit{n}}_{\mathit{t}}^{-}$	−0.694***
	(0.144)
Constant	−3.807***
	(1.201)
log likelihood	−1192.222
Cointegration F-test	209.50** [0.0507]
Long-run asymmetry Wald test	586.62***[0.0000]
Short-run asymmetry Wald test	5.26**[0.0218]

Note: Significance level (***p < 0.001, **p < 0.05, *p < 0.01). Standard error values are reported in brackets (), while p-values are reported in parentheses [] for long and short-run asymmetry. PMG method is employed for the estimation of the model.