Black and Hispanic people and people with low incomes are more likely to live in areas at high risk of flooding from natural disasters than white and Asian people, according to a new study led by the University of Arizona.

The study also found that certain reforms to the federal government's widely used flood insurance program could be disproportionately burdensome to the same groups of people.

The findings come amid a record-breaking 2020 Atlantic hurricane season that has pummeled the southeastern United States. Hurricane Sally, which made landfall on Sept. 16, devastated parts of Alabama and Florida. Two days later, forecasters at the National Hurricane Center ran out of the 21 storm names, following the Latin alphabet, that it had decided on for this year. Scientists turned to the Greek alphabet, for only the second time ever, to name the following storms.

Image

The study will be published in the November issue of the Journal of Environmental Economics and Management. Laura Bakkensen, an associate professor in the School of Government and Public Policy in the College of Social and Behavioral Sciences, was the paper's lead author. Bakkensen, an environmental economist, researches policy responses to natural disasters.

"We see a lot of news reports where sometimes poorer or minority-rich communities may feel greater losses," Bakkensen said. "Understanding whether that's happening and what might be driving that is really what got me interested in this."

A Database of Homebuyers' Choices

Bakkensen and her colleague Lala Ma at the University of Kentucky looked at four years of property sales data for nearly 50,000 homes in south Florida. The area has many properties prone to flooding, which made it a strong case study, Bakkensen said.

Bakkensen's study relied on individual property sales data linked with data on homebuyer ethnicity and income, providing a much more granular picture than many previous studies. This ultimately allowed researchers to get a clear idea of what factors played into homeowners' different choices about where to buy a house – or where not to buy a house.

"Flood risk house-by-house may be very different. My neighbor across the street from me might be at low risk but maybe I'm at a lower elevation and maybe I'm at high risk," Bakkensen said. "We wanted to be really careful to disentangle what is really flood risk versus a lot of other things that might correlate with flood risk in the data," such as proximity to the coastline, which is typically considered an amenity.

The key question from the data, Bakkensen said, became: How much would somebody be willing or able to pay to avoid living in a high flood-risk zone?

The answer was different depending on homebuyers' ethnicities, researchers found. The study reports that white and Asian people with low incomes would be willing or able to pay about $710 per year on average to avoid living in an area at high risk of flooding. Because Asian people accounted for only 2.5% of all the property sales data, they could not be separated as their own category. They were combined with white people because the two groups' average incomes were similar in the study's sample.

Among Black people, that number was around $500. Hispanic people, on average, were willing or able to pay an additional $618.

Predictably, the study found that the amounts homebuyers were willing or able to pay increased with their income levels.

The study notes that other factors beyond income – such as access to information about homes' flood risk – could explain why certain groups are able or willing to pay more than others to avoid flood risk. But determining that would require further research and additional data, Bakkensen said.

Reforming Flood Insurance Through Simulations

The study also sought to find ways to improve the National Flood Insurance Program, which has long been a target of reform by both political parties in Congress, Bakkensen said.

Under federal law, most lenders require flood insurance for homebuyers who are buying properties in areas at high risk for flooding. Due to a small market of private flood insurers, the federal program is the largest and most widely used, Bakkensen said.

But the seemingly competing missions at the core of the program – provide flood insurance and make it affordable for homebuyers – have made it difficult to fund, Bakkensen said. The program regularly borrows money from the Treasury Department to cover the cost of payouts.

Using the housing data and model results, Bakkensen and Ma ran simulations that predicted how certain proposals to reform the program could affect homebuyers. In one case, the researchers simulated removing two subsidy programs that make the insurance more affordable for homeowners.

They found the change would have disparately negative effects on homeowners of certain ethnicities. Removing the subsidies would cost white and Asian homeowners with low incomes only about 0.45% of their average annual incomes, compared to 0.55% of the average annual incomes for Black homeowners and 0.7% for Hispanic homeowners.

As homeowners' incomes increase, the cost of losing the subsidies took up less of their annual income, but the disparities for people of different ethnicities remain.

Another simulation took aim at the National Flood Insurance Program's flood risk maps, which are used by homeowners and officials to determine the flood risk of neighborhoods and homes in large sections of the U.S. But many of the maps are outdated, meaning the flood risk information in those maps is unreliable.

On average, Bakkensen and Ma found that updated maps provided significant value to homeowners, especially to lower income homeowners as a fraction of their income, and the researchers were able to translate those advantages into dollar figures. But the value fluctuated based on ethnicity: For white and Asian homeowners, revised maps provided an annual value of $144; that value was $70 for Hispanic homeowners and $60 for Black homeowners.

That finding alone has already raised eyebrows at the National Flood Insurance Program, Bakkensen said, where program staff have been trying to get an idea of the value of updating the maps.

The disproportionate costs of reforming the program to Black and Hispanic people and people with low incomes, Bakkensen said, were among the findings that surprised her the most. She had no pre-conceived notions, she added, about whether she would find disparities in reform outcomes.

"It was quite a striking finding, and I think it was an important one," Bakkensen said. "It shows that even with good-intentioned policy reform, we still have to be careful and do our due diligence to make sure there aren't unintended consequences that could cause harm to communities that may already experience a lot of harm from flooding and other natural disasters."

Bakkensen plans to expand her research to other areas of the country, as well as to renters, who were not included in this study. The new findings help illustrate just how differently certain groups of people experience natural disasters, she said.

"There's a lot of complexity to who's living in harm's way, and that's why we oftentimes see very different losses, or it can be an important contributing factor," Bakkensen added. "That's quite relevant for a lot of the hurricanes we're seeing this season."