This paper uses a Markov switching model (MSM) to decompose Macao's tourism cycle into high and low growth states (HGS, LGS) for the period of 2005Q2-2017Q2. The likelihood of the cycle maintaining HGS is 93% but the risk of staying in LGS is 80%. The Macao cycle is favorably asymmetric, with HGS (14.7 quarters) lasting much longer than LGS (5.1 quarters). Further, the paper combines structural regressions with the MSM to identify determinants of the Macao cycle, with useful policy implications derived from the regression results. We find that Macao's tourism cycle is heavily affected by Mainland China's business cycle and other external factors. Additionally, outward-looking marketing, albeit very costly, is found to be effective for keeping the local cycle in HGS.