Table 4. Results of short-run and long-run coefficients from ARDL(1,1,2,2,2,1) model.

Dependent variable is lnY
45 observations from 1970 to 2014 were used for the estimation
	Variable	Coefficient	Standard error	t-Ratio[Prob]
Long-run	lnM	0.2143***	0.0664	3.23[0.003]
coefficients	lnF	0.1969**	0.0789	2.42[0.011]
	lnA	-0.0384	0.4685	-0.08[0.935]
	lnP	-0.5536*	0.2890	-1.92[0.066]
	lnT	-0.6976	0.5071	-1.38[0.180]
Short-run	ΔlnM	0.1974	0.2427	0.81[0.423]
coefficients	ΔlnF	-0.2505**	0.1229	-2.22[0.035]
	ΔlnA	1.9851***	0.5950	3.34[0.002]
	ΔlnP	0.3071*	0.1689	1.82[0.080]
	ΔlnT	0.1608	0.2692	0.60[0.555]
Error correction coefficient	ECMt-1	-0.8167***	0.1639	-4.98[0.000]
	Cons	3.9876**	1.6782	2.38[0.025]
	Log likelihood = 111.8696 R-squared = 0.7961 VIF = 34.30		Adj R-squared = 0.6942 Root MSE = 0.0222

***Significance at 1% level;

** Significance at 5% level;

* Significance at 10% level