Abstract Many interesting issues are posed by synchronization of cycles. In this paper, we define synchronization and show how the degree of synchronization can be measured. We propose heteroscedasticity and serial correlation robust tests of the hypotheses that cycles are either unsynchronized or perfectly synchronized. Tests of synchronization are performed using data on industrial production, on monthly stock indices and on series that are used to construct the reference cycle for the United States. An algorithm is developed to extract a common cycle. It is used to extract the reference cycle for the United States and common cycles in stock prices and European industrial production.