This paper studies the impact of public debt-to-GDP ratios on the economic growth rate of 21 industrial countries during the 1990-2014 period. Thereby the aim is to determine the optimal debt-to-GDP ratio in terms of maximised economic growth. The debt-growth relation has been estimated by regression analyses, while taking alternative growth determinants, non-linearity and potential endogeneity issues into consideration. The findings confirm concavity of the debt-growth relation and the existence of upperbound values. The reported results indicate that moderate public debt levels stimulate economic growth, while diminished economic growth characterises debt ratios beyond the determined threshold. The most-encompassing instrumental variable specification indicates that the optimal debt ratio is about 129.55%, which is a higher public debt ratio than reported in previous studies. The conducted robustness checks show that the non-linear debt-growth relation prevails for alternative control variables, dependent variable, country sample and incorporated polynomial. Finally, the empirical analyses reported that the interaction with the real interest rate is a channel through which the public debt ratio reduces the economic growth rate. In the latter case, the inclusion of several interaction terms reduces the optimal debt-to-GDP ratio to about 100%.