Various kinds of shocks can influence tourism demand cycles. However, previous studies have typically concentrated within economic frameworks in explaining variations of tourism demand cycles, ignoring the possibility of different shocks influencing tourism demand cycles. This paper analyses various shocks in explaining the variation of tourism demand cycles to Bali on different source market countries. Monthly tourist arrivals data from January 2002 to January 2020 are decomposed into cyclical and irregular components. The cyclical patterns represent tourism demand cycles, and shock events are represented by irregular component patterns of international tourist arrivals data series. This study employs an unobserved component model (UCM) procedure to decompose the data, unit root tests to examine data stationarity, and logistic regression to analyse the relationship between shock effects on tourism demand cycles. The unit root test result indicates that shock effects on tourism demand cycles are not permanent. However, the persistence and magnitude of shock effects vary between source market countries. Shock effects could be persistent and influence the negative phase of the cycles for several countries. The persistence and magnitude of shock effects are the highest for tourist arrivals from the United States, India, and Germany. The results provide theoretical and managerial implications.