procurement before and after the floods. As discussed in Section 2, firms in not only flooded regions but
also non-flooded regions saw a period of no production because of the flooding. Experience of such direct
or indirect impact may raise firms’ subjective risks of local procurement, and thus persistently lower the
local procurement share. Furthermore, firms in flooded regions suffered physical damage to production
facilities such as capital goods. Such damage could raise the subjective risks of local procurement even
higher. To examine these questions, we investigate the share of local procurement before and after the
floods and compare changes in local procurement at affected and unaffected affiliates. This enables us to
understand which Japanese MNEs changed their subjective risks in the local market with a huge agglomeration
of suppliers.