Over the recent months, several initiatives have taken place to develop macro-prudential regulation in order to prevent systemic risk and the built-up of financial imbalances. Crucial to the success of such policy is the ability of the macro-prudential authority to identify in due time such imbalances, generally featured by asset-price boom-bust cycles. In this paper, we investigate the possibility of detecting asset-price booms according to alternative identification strategies and assess their robustness. We infer the probability that an asset-price boom turns into an asset-price bust. In addition, we try to disentangle costless or low-cost from asset-price booms. We find some evidence that house price booms are more likely to turn into recession than stock price booms. Resorting both to a non-parametric approach and a discrete-choice (logit) model, we analyze the ability of a set of indicators to robustly explain asset-price booms. According to our results, real long-term interest rates, total investment, real credit and real stock prices tend to increase the probability of a housing-price boom, whereas real GDP and house prices tend to increase the probability of a stock-price boom. Regarding the latter, credit variables tend to play a less convincing role. From this perspective, we specify the scope of macro-prudential regulation as a set of tools aiming at avoiding costly asset-price booms. In doing so, we try both to make the case for state-contingent macro-prudential regulations and to set out clear delineation between monetary and financial stability objectives.