In recent years, cost-benefit analysis(CBA) has played an important role in disaster risk reduction(DRR)investment decisions, and now increasing attention is being paid to its application in developing countries. This article discusses government investment choices in DRR against typhoon disasters in Shenzhen, China. While the existing literature mainly focuses on disaster mitigation measures such as structural retrofitting, this study proposes a holistic framework of DRR investments in which structural(windproof retrofitting) and financial(insurance premium subsidies and post-disaster relief) are all taken into account.In particular, intermeasure spillover effects are measured and used in CBA. The results show that insurance premium subsidies yield the highest benefit-cost ratio and should be prioritized in investment. Windproof retrofitting comes in second place in terms of the benefit-cost ratio and can be considered when there is a sufficient budget. These results further confirm the need of a holistic review of government DRR investments to derive policy recommendations, while challenges remain in relation to the probabilistic modeling capacity to support CBA. In recent years, cost-benefit analysis (CBA) has played an important role in disaster risk reduction (DRR) investment decisions, and now increasing attention is being paid to its application in developing countries. This article discusses government investment choices in DRR against typhoon disasters in Shenzhen, China. While the existing literature mainly focuses on disaster mitigation measures such as structural retrofitting, this study Propose a holistic framework of DRR investments in which structural (windproof retrofitting) and financial (insurance premiums and post-disaster relief) are all taken into account.In particular, intermeasure spillover effects are measured and used in CBA. The results show that insurance premium subsidies yield the highest benefit-cost ratio and should be prioritized in investment. Windproof retrofitting comes in second place in terms of the benefit -cost ratio and can be considered when there is a sufficient budget. These results further confirm t he need of a holistic review of government DRR investments to derive policy recommendations, while challenges remain in in to the probabilistic modeling capacity to support CBA.