Podcast
Important Information:
How Banks Are Adapting to Changing Flood Risks
Important Information:
Toan Phan reviews the risks faced by banks from the increased frequency and intensity of floods and other extreme weather events, and how banks and financial markets have responded to those risks. Phan is a senior economist at the Federal Reserve Bank of Richmond.
Related Links
Transcript
Tim Sablik: My guest today is Toan Phan. Toan is a senior economist in the Research department at the Richmond Fed. Toan, welcome back to the show.
Toan Phan: It's great to be back.
Sablik: It's definitely good to have you back on to talk about your research related to the risks that climate-related disasters such as storms and flooding pose for banks and the broader financial system.
The damage that major storms can cause is certainly very top of mind now for the Richmond Fed and many communities in our District. We're recording this episode on October 9 in the wake of Hurricane Helene and watching as Hurricane Milton approaches the Florida coast. The priority now, of course, is the safety and immediate recovery of communities in the path of the hurricanes. But thinking longer term, the economic impact of Hurricane Helene is expected to be substantial, and we don't yet know what the damage from Milton will be.
Toan, could you start by explaining the kinds of financial risks that major storms pose for households and banks?
Phan: The home is the most important asset for most American families. On the liability side, it is the most important liability of most families.
This is a question of how disasters — particularly hurricanes and flooding — affect mortgage performance. I've used some very large databases to look at how loans perform before and after hurricanes. For example, we look at the 2017 hurricane season, which was very dramatic. We have Hurricane Maria hitting Puerto Rico, Hurricane Harvey hitting Texas, and Irma hitting Florida. We've seen in our own data that these hurricanes have been causally linked to a dramatic increase in the risks and actual default rates of households' mortgages. So, we can see that hurricanes do have significant effects on the financial health of American families.
Sablik: One of the things that I've learned from reading your work and talking with you is how difficult it is to model and predict climate- and weather-related risks. In the case of flooding, for example, how do the ways that we measure and insure against flood risk affect the financial costs borne by households and banks?
Phan: Let's take a step back. The majority — I would say 95 percent, 96 percent — of all the flood insurance in the U.S. is provided by the federal government through a program called the National Flood Insurance Program [NFIP]. If you live in an area that is officially classified as a flood risk and if you're borrowing with a commercial mortgage that will be protected by the government-sponsored enterprises, then you'll be required to buy flood insurance.
This requirement is based on the official floodplain maps. Tim, if you live somewhere near the river or near the coast and historically your house or your location has experienced a lot of flooding, then you're likely to be classified [as in] the high-risk area. Then, you'll be required to buy flood insurance at a relatively subsidized rate provided by the NFIP. Now, if Tim's brother lives outside of the floodplain map, then there's no requirement, okay? You don't have to buy flood insurance. In fact, outside of the official floodplain maps, flood insurance coverage is very close to zero — 1 or 2 percent.
The official floodplain map is what really determines who has and who doesn't have flood insurance. So the question is how good are the official floodplain maps in measuring flood risk? If you think about historical flood risk, I think the official floodplain maps do a reasonable job. However, if you're thinking about the future flood risk, say 30 years over the duration of the loan, the flood risk is really changing because of factors like climate change.
Flood risks can come from different sources. One is coastal flood risk. With more energy in the ocean and with the rising sea level, the coastal properties will face a higher risk of storm surges. But the story doesn't end there because now the hurricanes, as you've seen recently in Helene, bring a lot of rainfall inland ... Now we have much more energy, and there's more wind that can carry rain and also flooding coming from the river, so the river overflows. So, the rainfall contributes another dimension of flood risk.
The official floodplain maps do an okay job capturing coastal flood risk. However, it's really the inland flood risk I think is underestimated by the official floodplain maps.
So, a group of climatologists and scientists have built an organization called the First Street Foundation to provide data on future flood risk projections. Many scholars in academia have been using it. I've used it myself. If you consider First Street Foundation to be like a projection of future flood risk, then we compare this with historical floodplain maps. You see there's a lot of differences for inland flood risk.
Think of Asheville, for example, North Carolina, a beautiful town [at] a high elevation in the mountains. I was very surprised when I see that it was underwater with such amount of torrential rain. The majority of Asheville is not classified as at high flood risk.
Sablik: Sure.
Phan: In fact, I think lower than 1 percent of households in North Carolina and Asheville in particular have flood insurance.
Sablik: Yeah, you can certainly imagine how people living in communities that are in areas that are not historically prone to flooding and might be underinsured [and] the financial risks that that poses to them.
On the flip side, the mortgages which are issued by banks in these areas might be exposed to greater risk as well. You recently wrote a paper with some co-authors studying how banks responded to Hurricane Harvey in 2017.
Phan: The 2017 hurricane season was quite important, as I mentioned to you. Some really costly hurricanes were really big signals for banks to think, oh, the flood risk is a serious event. In particular, [with] Hurricane Harvey, a lot of the actual flooding was inland. I think it's a wake-up call for society that inland flood risk is serious.
As a side note, inland flood risk depends on several factors. One of them is how pervious the ground is. In my previous work, I've shown how properties and houses in redlined areas tend to have lower perviousness. Think of parking lots rather than lawns.
I think that's one of the things that happened in Texas for Hurricane Harvey. A lot of places were flooded, in part, because they didn't have enough absorption of the ground, leading to a lot of inland flooding. That led to a lot of increases in the default rates for mortgages, as I mentioned to you earlier.
Sablik: What did you find when you looked at the banks that were in the path of the hurricane [and] how they adapted, versus those that weren't directly affected?
Phan: In this project that I had with my colleagues Dasol Kim and Luke Olson at the Office of Financial Research, we looked under the hood of a large database of banks [at] how banks managed flood risk — inland flood risk, in particular — before and after the 2017 hurricane season. We found that banks that had a lot of lending activities in areas that were directly hit by the hurricanes changed internal risk models a lot. Banks have internal risk models so they can evaluate the risk of default.
What we found is that after the hurricane, banks significantly revise the default risk for houses that have higher future flood risk, and that's true even if that house is not directly affected by the hurricane. In other words, banks seem to be learning. They woke up after 2017, in a way, and realized the importance of inland flood risk and started to price in this inland flood risk on the loans, even if the houses were not directly affected by the hurricane in 2017 itself.
Sablik: Yeah, they're learning. They're updating their beliefs.
One of the other things that was very interesting I found in your paper was that there are also other factors that affect how the banks are responding and updating their risk beliefs in the future. One of those factors that you study is competition. How does the competitive environment where these banks operate affect their decisions and how they update their risks or their expectations for risks in the future?
Phan: It's really interesting because when it comes to unmapped flood risk, there's not much regulation. The regulation is if a house is inside the official floodplain maps, then there are requirements that you have to buy flood insurance if you're borrowing with a conventional mortgage that's guaranteed by the government-sponsored enterprises. Outside of these regulated areas, there is very little regulation. The market forces have a lot of room to play.
I think that was a motivation for this, for our question, "How do competitive forces apply in the absence of regulatory oversight?" What we find was that in places where the banks are more competitive — big banks have other big banks in the neighborhood providing the same kind of loans — then we see that banks tend to manage risk less. In other words, they're less likely to manage future flood risks when they face competitive forces.
What kind of management are we talking about here? Let me be clear, we are talking about home equity lines of credit. We're talking about second lien loans here. These are loans that banks cannot easily sell to the market through securitizations, unlike mortgages. For home equity lines of credit [and] home equity loans in general, banks generally have to hold onto these loans until maturity. They cannot sell them off, so they have a pretty strong incentive to make sure that their loans are performing okay. They're not going to underwrite risky loans.
What we found was after 2017, banks were managing this future flood risk by reducing the bank's exposure, by reducing the lending to at-risk places. However, this risk management is weakened when there's more competition. In markets where there is more competition in the home equity lines of credit, banks are less likely to reduce risky lending.
Sablik: Is the idea there that banks are concerned about losing out to competitors if they're being too risk averse?
Phan: That's one dimension. When there's more competition, you have fewer margins of adjustment. For example, you might lose your market share to your competitors. So, when they have less margin for adjustment because you have more competitive factors, you're less likely to adapt.
Sablik: Gotcha.
Thinking about the findings from your paper, and then more generally what we were talking about with the floodplain maps and flood insurance, what are the lessons that come out of that for policymakers and for financial regulators, in the case of the banks?
Phan: I've been thinking about this a lot. Our findings suggest that banks are adapting. Banks are learning, they're revising the internal risk models to evaluate flood risk better, and they are reducing their exposure to flood risk by reducing lending.
Now our findings are about home equity lines of credit. However, when you look into mortgages in other papers, people have found that banks are adapting in a different way. Banks are adapting by securitizing and selling off loans that are at higher risk of future flooding to the government-sponsored enterprises. So, they're adapting by shifting the risk to someone else, in this case the government. The GSEs, Fannie Mae and Freddie Mac, have been under government conservatorship since the financial crisis. So, the bottom line is banks are adapting, but how they adapt depends a lot on institutional details.
A lot of initiatives right now when it comes to managing emerging risks like climate risk, a lot of them are focusing on micro-prudential policies. The focus is on individual institutions — are their policies foolproof when it comes to the future disaster risk? So, the micro-prudential policies focus on individual banks' stability.
Our findings suggest that we also should think of macro-prudential policy because market forces like competition or concentration matters, right? When it comes to the re-shifting of risky mortgages to the GSEs, I think that brings back this concern that we had before the financial crisis — who, in the end, is going to hold the risk? Why are the GSEs buying these risky mortgages? The GSE policies, when it comes to guaranteed loans and the fees they charge for guaranteeing these mortgages, tend to not reflect underlying flood risk. They tend to rely only on, generally, the official floodplain maps that I mentioned to you earlier, which underestimate a lot of this future flood risk.
Sablik: Yeah, certainly a lot to think about and keep in mind.
Thinking in the more long-term perspective, is there anything in particular related to this research that you plan to expand in the immediate future?
Phan: I think there are a lot of open questions remaining about how firms or households choose or choose not to insure, or choose to under-insure or not. This average matters for financial stability because underinsured loans tend to experience a lot higher default risk after disasters. That's something that we've seen over and over, time again, in many research papers after many different disasters.
The bigger questions are to the extent to which the National Flood Insurance Program, by subsidizing flood insurance at relatively lower insurance premium rate, is creating a disincentive for families to not move away from risky areas. Also, is there a way the subsidized premium might create a friction against housing prices in the market to fully reflect underlying flood risk? These are some of the questions that many people are thinking about, and this is something that my co-authors are also trying to understand.
Sablik: Yeah, very interesting.
Toan, thanks, as always, for joining me on the show to talk about it all.
Phan: Thank you for having me here, Tim.
