1. The AMG estimates for Belgium show UR coefficient of -92.256. Given all variables are in natural logs, this implies a 1% increase in urbanization reduces LCF by 92%, economically implausible. Similar issues appear for Portugal (GDP coefficient -214.262) and Switzerland (UR coefficient -50.492). These likely reflect model misspecification or coding errors.
2. Table 8 shows LnCF is stationary at level I(0), yet the Westerlund cointegration test (Table 9) assumes the dependent variable is I(1). Using I(0) dependent variable with I(1) regressors invalidates the cointegration framework.
3. “Lngdp” appears twice with different values (5260.124 and 5245.256) while “Lngdp2” is listed but never defined in the model. The model in Equation 1 includes no squared GDP term.
4. The authors state “UR increases LCF in France, Hungary, Norway, Slovak Republic” (p.10). However, Table 10 shows France’s UR coefficient is actually negative in the CCE estimation (-5.349, insignificant) and only positive in AMG—but they don’t specify which estimator they’re reporting.