In the Netherlands one of the most important methods to determine the level of flood protection is a cost-benefit analysis. In this analysis the costs (investment and maintenance costs of dikes) are balanced with the benefits (a reduction of flood risk). Risk is defined as the product of probability and consequence. This approach requires a good estimate of the consequences of floods. A large contributor to the total costs of a flood are the losses due to business interruption. This is defined as forgone value added that is not created due to the fact that firms have to stop production. The causes of the production stop considered in this thesis are material damage to the production facilities of the considered firms and both forward and backward effects in the supply chain. The latter effects include the interruption of firms which are not flooded themselves, unlike their suppliers or buyers. In the Netherlands the software package HIS-SSM (Hoogwater Informatie Systeem – Schade en Slachtoffer Module) is used to determine the consequences of possible floods. This model uses a damage function approach to determine the losses due to business interruption, in which the duration of the business interruption depends on the water depth only. As a result of this assumption, the losses due to business interruption expressed as a share of the material damage range between 1.5 and 5%, independent of the scale of the flood according to HIS-SSM. An analysis of actual floods shows that these figures are between 30% and 125% for large-scale floods. The considered floods for these figures are hurricane Katrina in 2005, hurricane Sandy in 2012, the tsunami in Japan in 2011 and the river floods in Thailand in 2011. Therefore, the losses due to business interruption as calculated by HIS-SSM are expected to be an underestimation of the actual losses due to business interruption. For this reason the ARIO (Adaptive Regional Input-Output) model has been used to determine the losses due to business interruption as well. The ARIO model has only been used for academic purposes so far. Both models have been applied to three cases in the Netherlands, which comprise the flooding of the province of South-Holland, Groningen and the small town of Arcen. The results of the case studies show that the estimates of the losses due to business interruption as calculated by the ARIO model are more in line with figures from actual floods. Both HIS-SSM and the ARIO model have some drawbacks however. The ARIO model appears to be able to determine the losses due to business interruption more accurately than HIS-SSM. Many important processes in the aftermath of a flood are not considered in both models however, such as flood duration and substitution effects. In both models the consequences of flooding of unique firms or infrastructure are underestimated. This is a result of the sector approach in both models, in which all firms are assigned to a certain sector. The transformation of material damage to production capacity has a large influence on the results of the ARIO model. The costs of permanent relocation of firms to foreign countries is also not taken into account in these models, while a simple calculation shows that these might contribute significantly to the total costs of a flood. Some assumptions in this calculation are based on interviews conducted with a member of the Dutch parliament and a professor at Delft University of Technology. Measures to reduce the losses due to business interruption should also be considered in the Dutch flood protection policy, after a cost-benefit analysis has been performed for these measures.