*3.2. Cointegration*

Accordingly, this sub-section describes the type of cointegration analysis employed in each of the sampled studies. First, the study by Menegaki and Tugcu [33] in page 31 used a panel cointegration procedure developed by Westerlund [49] which considers the cross-sectional dependence that has been previously acknowledged. Second, the study by Menegaki et al. [34] in page 1261–1262 employed the ARDL cointegration framework which directly hosts both short run and long run relationships. Third, the study by Menegaki and Tugcu [35] in page 897 (Table 4 in the referenced paper) uses the panel ARDL model with a Pooled Mean Group (PMG) estimator. The advantage of this estimator is that it allows the intercepts, the short run coefficients, and error variances to differ across groups of countries, but it constraints the longrun coefficients to be the same. Fourth, the study by Menegaki and Tiwari [31] in pages 501–502 applies the Pedroni cointegration test [50] and based on evidence from the Hausman test they have decided in favour of a dynamic fixed effects model to depict the cointegration relationship estimated with a Generalized Method of Moments (GMM), which revealed that there was no problem of autocorrelation. They have also employed a quantile regression to corroborate the previous results. Fifth, the study by Menegaki and Tugcu [32] in page 83 has also used the Westerlund [49], whereby the underlying idea is to test for the absence of cointegration by determining whether the individual panel members are error correcting. Over the long run, cointegration is employed to investigate whether the variables move along the same path. The confirmation of cointegration also indicates the existence of a causality relationship at least in one direction of the relationship. Menegaki and Tugcu [30] in page 157 have used the Pedroni [51] cointegration with seven test statistics. Four out of the seven statistics are estimated based on pooled data across countries, and three out of the seven are based on averages of the individual autoregressive coefficients for each country.
