Estimates of the Difference in Average Flood Damages Due to Having a Flood Insurance Policy (in Euros)
| Contents Damage | Building Damage | |
|---|---|---|
| Comparison of mean flood damage suffered by households with and without flood insurance | 3,458*** (928) | 9,514*** (2,537) |
| ATT based on PSM using as matching method: | ||
| Nearest-neighbor matching | 2,126 (2,241) | 6,337 (5,852) |
| Radius matching | 1619 (1,684) | 7,016 (4,490) |
| Stratification matching | 1,395 (1,874) | 6,261 (4,668) |
| Kernel matching (Gaussian) | 1,832 (2,075) | 6,196 (4,013) |
| Kernel matching (Epanechnikov) | 1,583 (1,978) | 7,266* (7,266) |
| Average ATT estimate | 1,711 | 6,615 |
| Number of matches | 270 | 255 |
| Variables described in Appendix A | 3–32,34–40 | 3–8,10–32,34–40 |
Note: The numbers in parentheses are standard errors. Where analytical standard errors are not available, they have been calculated via bootstrapping with 2,000 repetitions. The average treatment effect on the treated (ATT) is estimated using propensity score matching (PSM) with different matching methods. The ATT estimates above have been rounded to the nearest whole euro. For a list of variables used in the propensity score function refer to Appendix A.
↵*** Statistical significance at the 1% level.