This paper investigates whether the use of the Hamilton regression (HR) filter significantly modifies the business growth cycles of GDP, as compared to those extracted with the double Hodrick-Prescott (HP) filter in the OECD System of Composite Leading Indicators (CLIs). Hamilton (2017) recommends using a regression filter to overcome some of the drawbacks of the HP filter, which includes the presence of spurious cycles, the end-of-sample bias and ad-hoc assumptions on the smoothing parameters. In our analysis, we assess whether the use of the HR filter significantly modifies the number and the dating of turning points. We then measure the degree of synchronisation of the resulting cycles as well as the main features of their phases. Finally, we compare the behaviour of the two filters at the end-of-sample in a quasi real-time framework. Results suggest that, for most of the OECD and BRIICS countries, the reduction in the number of turning points, and hence in the number of cycles, is actually negligible. Moreover, our analysis reveals that the chronology of turning points and the end-of-sample direction of the HR filter is significantly more volatile than that of the HP. We conclude that the use of the HR filter for recurrent forecasting of turning points and end-of-sample direction of the growth cycle may be affected by large revisions, lessening the credibility of the message to policymakers in the long run.