Capital mobility in Latin American and Caribbean countries: new evidence from dynamic common correlated effects panel data modeling

Financial Innovation

Table 3 Cointegration tests

1. The Westerlund panel cointegration test (2007) results
Test	Value	Z-Value	P-Value	Robust P-Value
G_τ	2.919	−3.044	0.001	0.000
G_α	−15.849	−2.524	0.006	0.010
P_τ	−13.224	−4.139	0.000	0.000
P_α	−15.906	−5.086	0.000	0.000
2. Gengenbach et al. (2016) panel and individual country error correction test results
Country	α_yi	T α_yi ^a
Argentina	−0.522	−4.02**
Bolivia	−0.345	−3.234
Brazil	−0.417	−3.850***
Barbados	−0.283	−3.08
Chile	−0.207	−2.08
Colombia	−0.529	−4.702*
Costa Rica	−0.674	−5.266*
Dominican Republic	−0.437	−3.549***
Ecuador	−0.201	−1.919
El Salvador	−0.729	−4.439*
Gautemala	0.215	−2.319
Guyana	−0.484	−4.195**
Honduras	−0.429	−3.305
Mexico	−0.478	− 3.624***
Jamaica	−0.393	−4.383**
Nicaragua	−1.229	−5.201*
Paraguay	−0.240	−2.692
Peru	−0.374	−4.057**
Trinidad and Tobago	−0.398	−3.228
Uruguay	−0.735	−5.229*
Panel (ECT for Investment)	−0.466	−3.719^*
LR Savings Retention Coefficient	0.240	2.54**
CD	−3.74 (0.000)^b	RMSE =2.80

α_yi = the error-correction term for the i_th country_. Null hypothesis is no cointegration. The robust p-values are based on 200 bootstrap replications (with lags (1), leads (1), lrwindow (1). The results are robust with other orders of lags, leads and lrwindow and 400 bootstrap replications. α_yi denotes the error-correction term
^a The lag length of 4, based on [4(T/100)^2/9] (See Gengenbach et al. 2016). *, ** and *** denote statistical significance at the 1%, 5% and 10% levels of significance, respectively
^b The observed p-value in the parenthesis. Critical values are taken from Tables 1 and 3, with m = 1 and no deterministic term (Gengenbach et al. 2008)