Not everybody reacts to floods and not everybody does it the same way: Adaptation of private flood protection after a major flood event

In June 2013, a major flood hit great parts of Germany, causing economic damage of several billion €. This event illustrated the increasing severity and frequency of flood events in Europe, which will be further aggravated... [ view full abstract ]

In June 2013, a major flood hit great parts of Germany, causing economic damage of several billion €. This event illustrated the increasing severity and frequency of flood events in Europe, which will be further aggravated by climate change in the future. As part of a comprehensive flood management and climate adaptation, policymakers stress the importance of technical and behavioural measures to reduce or avoid flood damage at the household level. It was repeatedly argued and shown empirically that personal experience of flood damage is a major driving force for the motivation to engage in these proactive flood protection measures. In this paper, I add to this literature in several dimensions. First, in contrast to many existing studies, I use a large sample from the general population of Germany (more than 4,000 households), as opposed to focussing on a specific flood-prone area. Second, by using longitudinal survey data from 2012 and 2014, the causal effect of experiencing the flood of 2013 can be assessed by a difference-in-differences estimator. This procedure deploys the difference in the time trends of flood protection between households experiencing the flood and non-affected households. Third, the results provide detailed insights into the heterogeneities of private reactions to flood events, such as typical characteristics of reacting households, and which kinds of measures they predominantly employ.
The results demonstrate that the flood of 2013 indeed had a significant effect on the willingness to engage in flood protection measures after the flood. The effect is strongest amongst those households who report themselves as being damaged or evacuated during the flood and those households living in counties with a high frequency of insurance claims following the event. If experienced households are identified by living in a county which initiated flood alert or by assisting other affected households, the treatment effect gets smaller and insignificant. Mostly the reacting households are homeowners, but tenants also increase their protection behaviour if they were damaged or evacuated during the flood. Furthermore, reacting households live in a flood prone area, and report they are flood insured. The highest increase can be observed for a relatively cheap, behavioural measure, namely shifting valuable assets to higher storeys. The results are relevant for the design of policies supporting private flood protection measures and better understanding the adaptive reaction of households to extreme weather events.