The Power of Data Transparency
More than 98% of residential flood insurance is administered through the National Flood Insurance Program (NFIP).(1) The program, which has over 1.5 million flood insurance policies that provide over $1.3 trillion in coverage, has seen over $31 billion in payouts over the last ten years alone (2007–2017).(2) The amount of claims due to flood disasters(3) has increased in recent years, with 2005, 2008, 2012, 2016, and 2017 being the highest on record and 2019 having the potential to be added to that list.(2)
In June 2019, the Federal Emergency Management Agency (FEMA) released 40 years worth of NFIP data, including nearly 50 million policy and payout records from the last ten years, over 2 million claims records from as far back as the 1970s, and files regarding FEMA’s reinsurance placement.(4) The release by FEMA aims to promote transparency, allow property owners to better understand their flood risk, and to aid the private insurance market in growing in order to offset the burden on the NFIP. The data paints a fair picture of where flood damage has occurred and the costs associated with that damage, but lacks the specificity necessary to determine accurate risk projections for homeowners. Nevertheless, with searchable details such as state and zip code, year of loss, and the dollar amounts for paid out claims, the data allows for a number of valuable insights.
Public access to its claims data could help fix the NFIP’s growing costs, fill gaps in coverage for flood-affected communities, and help cities better plan for disasters.
BENEFITS OF DATA RELEASE
Address Growing Costs. Billion dollar flood disasters in the U.S. have cost over $1 trillion in inflation adjusted dollars since 1980, with damage totals steadily increasing through 2018.(5) The increasing monetary amounts paid out for disaster recovery directly correlates to an increase in the number of extreme flooding events(6) and their growing intensity and destruction. Due to this, the NFIP has been unable to financially sustain itself while meeting the cost of flood insurance claims. Over time, these losses have accumulated, and even with Congress forgiving $16 billion worth of debt in 2017, the program still faces a $20.5 billion deficit.(7) The claims data released by FEMA could allow economists, flood experts, and researchers to better analyze the NFIP and lead to more informed policy decisions and reform of the program and insurance markets.
Fix Insurance Gaps. Analysis of the recently released FEMA data reveals that while homeowners in some states are paying significant annual premiums, they are claiming minimal amounts in damages. On the other hand, homeowners in some states are paying less in total premiums but are submitting claims and collecting large payouts for flood damage.(8) By looking at premiums verses payouts, policies can be adjusted to close the substantial insurance gap in the country. Properties in the underserved or unserved markets could be offered coverage through similar policies to what the NFIP currently provides at comparable or lower prices.
Aid City Planning. This data can be used to understand disaster trends, by examining repeated spikes in the number of claims for certain areas. This can help cities plan for flood insurance gaps and the financial implications that come with flood disasters. For every $1 spent on pre-disaster mitigation, communities save as much as $4-$6 on disaster recovery costs.(9) Having a clearer picture of flood risk and the scale of disasters an area is vulnerable to would allow municipalities to implement the proper adaptation measures that prevent flooding and withstand catastrophic events. As can be seen in examples in New York,(10) New Jersey,(11) Connecticut,(12) Massachusetts,(13) and others, the more information that cities have to plan with, the more successfully they can strategize for extreme flooding events.
THE WHOLE PICTURE
While this data release is a step in the right direction and offers great insight into the NFIP, it is presented in a way that may be difficult for the average person to understand or navigate, which creates a barrier to access for the general public. OpenFEMA, which houses the FEMA data, allows users to access a Data Dictionary to better understand what they’re reading,(4) but the information remains complex and in multiple file locations, and in files too large to import into consolidated Excel spreadsheets. To adequately access the data, users would need a proprietary tool.
Furthermore, redacting individual home addresses and releasing the information on a zip code level leaves information gaps and fails to give the full picture. Homes within each zip code may differ significantly in their risk exposure due to variations in a number of factors including elevation and distance from the coast.
FEMA has made it clear that the original purpose of its flood maps was not for risk communication, but to establish flood protection guidelines and aid in floodplain management.(14) The recent data release shows the need for improved mapping, to identify properties that may be at risk and uninsured, and further demonstrates the need for improved flood risk communication, through a user friendly tool.
Come 2020, FEMA will be implementing their Risk Rating 2.0, which is intended to improve the NFIP rating system by combining industry technology with NFIP mapping data.(15)
NECESSARY TRANSPARENCY
The transparency offered through FEMA’s NFIP data release is a big step. It can help both the public and private sectors to better understand flood risk and NFIP insurance claims on a broad level. However, while the data gives a glimpse at what areas are at greater or lesser risk, it does not paint the full picture. With the potential for significant variation in risk from home to home, providing claims data on a zip code level could create misperceptions of risk. Releasing the data on a property-lot level would significantly improve understanding of true risk.
Full access to the NFIP claims data on the most granular level would be just a fraction of the flood risk equation. To have the full picture, property owners and municipalities should understand flooding history beyond documented insurance claims. They need to understand the impact of nearby road and property flooding on home value, as well as the growing risk of tidal, storm surge, riverine and rainfall-driven flooding. This data should be concise and easily accessible in one central location.
One of the core principles of data science is the idea of transparency, both in the sense that data collection methods should be open to scrutiny, and that the results of data collection should be shared with the general population. Limited data or data published in a way that is not easily interpreted by those outside the science or data science community limits true understanding of flood risk and accurate risk projections. Without individual property data, and all data points, homeowners are prevented from knowing the full flood history of their home and from understanding their future risk of flooding.
Originally published at FirstStreet.org on July 23, 2019.
