Growth, bank credit, and inflation in Mexico: evidence from an ARDL-bounds testing approach

Latin American Economic Review

Table 3 Unit root test

Variables	ADF		PP		KPSS		Decision
Variables	Level	1st diff.	Level	1st diff.	Level	1st diff.	Decision
GDP	−2.706***	−4.794*	−2.552	−4.794*	0.807*	0.377***	I(0)/I(1)
GOV	−7.417*	−3.268**	−7.417*	−3.131**	0.751*	0.621**	I(0)/I(1)
INV	−1.834	−6.393*	−1.976	−6.393*	0.433***	0.310	I(1)
XP	−1.104	−4.398*	−0.995	−4.370*	0.811*	0.131	I(1)
INF	−1.624	−6.654*	−2.317	−5.841*	0.210	0.393***	I(1)
PC	−1.513	−3.218**	−1.339	−3.566**	0.495**	0.070	I(1)
BA	−3.847*	−6.330*	−2.728***	−6.552*	0.070	0.190	I(0)/I(1)
LL	−2.668***	−4.088*	−1.962	−3.993*	0.111	0.089	I(1)
FD	−2.867***	−4.972*	−1.917	−4.732*	0.139	0.084	I(1)
INF × PC	−5.023*	−5.411*	−5.035*	−28.574*	0.169	0.346	I(0)
INF × BA	−1.692	−6.929*	−2.371	−7.925*	0.224	0.500**	I(1)
INF × LL	−1.486	−7.014*	−2.185	−7.147*	0.235	0.247	I(1)
INF × FD	−1.615	−6.909*	−2.407	−8.553*	0.259	0.500**	I(1)

AIC was used for ADF to select the lag length; the maximum number of lags was set to five. Barlett–Kernel was used for PP and KPSS, as the spectral estimation method
*, ** and *** are statistically significant at 1, 5 and 10 %, respectively