Designing appropriate post-disaster emergency and recovery housing policies requires accurate estimation of the indirect or “flow” loss generated by the involuntary displacement of households to housing other than their own destroyed homes. We employed the stated choice method to measure residence choice following a hypothetical disaster in a procedure developed to estimate flow loss due to house destruction. This method was applied to households in the city of Nagaoka in Niigata Prefecture to estimate flow loss for a range of residence types and residence attributes following relocation. The results revealed that the relocation residence type itself (such as a shelter, temporary dwelling, rental housing, or one’s own home) had a substantial effect on residence choice. Regarding residence attributes, residential expenses such as rent, housing loans, and spaciousness had a significant effect on residence choice. Moreover, we found that respondents placed a very high value on living in their own home compared with other residence types, indicating that subsidizing the prompt repair and rebuilding of private homes is likely to be an effective recovery support policy. In addition, a cost–benefit analysis of our results revealed that rent subsidies are more efficient than the construction of public housing as a means of financially assisting economically constrained households after a disaster.