Does public flood aid crowd out private insurance?

How to cite

Garbarino, Nicola; Schickfus, Marie-Theres von; Neumeier, Florian; Möhrle, Sascha. 2021. "Does public flood aid crowd out private insurance?." AEA RCT Registry. AEARCTR-0008603. https://doi.org/10.1257/rct.8603-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Nicola Garbarino* , Marie-Theres von Schickfus , Florian Neumeier , Sascha Möhrle

* Primary investigator

Registry ID
AEARCTR-0008603
DOI
https://doi.org/10.1257/rct.8603-1.0
First registered
Last updated
Status
Completed
Trial dates
2021-11-06 – 2021-11-26
Countries
Germany ()
JEL codes
H84, G52, Q54
Registry record
http://www.socialscienceregistry.org/trials/8603

Abstract

Policymakers often do not want to promise flood aid ex-ante, to encourage households and firms in flood-prone areas to take up flood insurance. But ex-post, they feel compelled to provide aid to flood victims. This tension is visible in recent policy decisions in Germany, where commitments to stop providing aid were followed by promises to cover most of the damage suffered by the victims of heavy floods in the summer of 2021.

We perform an information treatment where we provide respondents with information about these policy decisions. The sample is split in three groups, which receive, respectively: 1) information about commitments to cut flood aid ("low aid"), 2) actual aid to flood victims ("high aid"), and 3) generic information about flooding world-wide.

We ask expectations of flood aid before and after the information treatment, to check whether expectations are updated. We then ask about two main outcomes: the willingness to pay for private flood insurance, and support for a policy that introduces mandatory flood insurance.

The starting hypotheses are:

1) the high (low) flood aid treatment decreases (increases) both willingness to pay for flood insurance and support for mandatory flood insurance;

2) the effect is concentrated among households that i) express a high concern about flood risk; and ii) update their expectations substantially as a result of the information treatment.

We conduct separate surveys of households (online, across Germany) and firms (both by phone and online, focussing in an area with high flood risk). This trial registration refers to the household survey. The firm survey will be registered separately.

Intervention

Information treatment on a sample of German households.
We provide the following information:
- Group 1 (active control group): % of natural disasters (worldwide) that are due floods or storms
- Group 2 (high flood aid information): households and firms in parts of Germany hit by heavy floods in July 2021 will receive flood aid up to 80% of the damage
- Group 3: (low flood aid information): a number of German states have recently committed to stop providing flood aid

Experimental design

- Sample split in three groups of similar size. Each group gets information about flooding.
- The control group receives generic information about the importance of flooding.
- The two treatment groups receive information about two recent policy decisions in Germany related to flooding. The first policy is the announcement by some German states that they plan to stop providing flood aid (with some exceptions). The second policy is the actual flood aid (80% of damage) promised to the victims of the heavy floods that hit parts of Germany in July 2021.

Design details

Randomization method
Randomization by computer, carried out by survey company (forsa)
Randomization unit
Households
Clusters
0
Observations
Households: 15,000
Arms
Circa 5,000 in each group: "high aid", "low aid" and control.
Primary outcomes
- Willingness to pay for flood insurance - Support for a public policy introducing mandatory flood insurance

Cited by

No citation data available for this registration yet.

Keywords