What Fannie Mae is doing about climate risk's impact on mortgages

June 19, 2023 4:00 AM

Climate risk can be tough to quantify, but doing so has grown in importance due to regulatory attention and increased difficulty obtaining property insurance coverage for it.

In response, Fannie Mae Chief Climate Officer Tim Judge is using his background in analytics and modeling to try to predict and address its impact on the government-sponsored enterprise's sizable mortgage portfolio. It's an important risk to manage for mortgage companies, homebuyers and taxpayers because Fannie buys a significant number of the loans made in the United States and is held in government conservatorship.

Judge has been working for almost two years on a long-term effort to assess the accuracy of internal and external data and modeling at Fannie. He identifies the largest risk as flooding but looks at many others that range from wildfires and the shift to renewable energy sources to derechos, a wind condition associated with thunderstorms.

Judge also has worked to stage community outreach efforts and surveys aimed at building consumer awareness about risks and sharing information about what they can do to make their homes more resilient. He has used multiple channels, including social media, to do this and also worked to ensure the distribution of information is in line with Fannie's equitable housing plan, which aims to reduce racial disparities in residential real estate.

Read on to learn more about Judge, the evolution of his role at Fannie Mae, and the property risk that mortgage companies and borrowers face. The excerpts that follow have been edited for clarity.

How did you come to take this job? What drew you to it?

I have a background in risk management modeling. What brought me to the job is about three years ago, we were looking at the flood risk at Fannie Mae and management wanted to get a better understanding of what our exposures were there. I thought the consideration was broader than that, and that we needed to be more holistic in terms of thinking about what the overall climate impact is to Fannie Mae. I brought a proposal to the management team and they said we should be looking at this as a broader issue. We've built out our climate impact team over time. It's a really impactful and very challenging area.

At this point, are there particular regulatory or other mandates that you are looking to fulfill ?

I'd say it's mostly internally driven. We certainly have a scorecard with FHFA, and FHFA cares about climate items but most of my roadmap and where we're going is driven by our internal focus on it. We are mindful of the SEC proposed rulemaking on climate disclosure. Certainly, once that gets finalized, that will have a big impact on our roadmap as well.

My understanding is the equitable housing plans call for some examination of climate impacts?

We put in some of our efforts relative to climate in what we're doing relative to housing equity. What we have there is us raising awareness at the community level and bringing climate analytics to different communities that may not have access to them otherwise. We are working with both Memphis and Baltimore to provide them with analytics on what those areas might look like now and in the future due to climate change. What's their exposure to heat? What's their exposure to flooding in the future? And we talk to them about what possible actions they can take.

I’ve heard some of the climate risk analytics are becoming less retrospective and a little more can be done in the way of projections. Is that true?

There are two different types of vendors modeling in this area. There are climate models, like Jupiter's, which are prospective. Firms like Jupiter and First Street are telling you what's going to happen in 2050 or 2100. Then there are catastrophic modeling firms who have been the ones that everybody's used to set insurance rates. It's interesting, one tells you what the statistical likelihood of a hurricane is next year. One tells you what the environment and the climate might look like in 30, 50 or 100 years.

One of our jobs is to take both those kinds of disciplines, put them together and figure out what's the best usage of both those types of disciplines in order to come up with how we should manage risk. Catastrophic firms are saying, okay, we can start to tell you what losses look like and damage looks like due to hurricanes in the future; and the climate firms are starting to say, we can tell you get more toward present day what your risk is. It's really an evolving space.

Are the insurance issues around wildfires in California and the ones in Canada part of what you're examining?

We always start with flood risk. Flood is the biggest risk right now. It also changes considerably with climate shifts. But people are starting to become more focused on a number of areas. We focus on some of what are called chronic climate risks. To us, the long term impacts of extreme heat and water scarcity are pretty big, but we are starting to focus some more short-term efforts on what I would call the availability and the affordability of insurance. It's always been an area of focus but now for us, it's probably a more short term area focus than historically.

Wildfires are starting to really impact availability of insurance and the cost. From the perspective of how we manage it, it's still every loan that we acquire has to have full fire insurance at replacement value, so it's not as an immediate concern from a loss perspective, so much as what's the impact to the consumer?

I would also tell you, what they call severe convective storms — think about tornadoes and derechos — those are actually starting to widen out the terms of the places they hit. That's another thing that we've started to focus on and the losses in those areas have also increased.

We have to be focused on if there's going to be some insurance pushback on some of that coverage as well in the longer-term.

How do you manage these risks with analytic tools?

We're still working on that. I don't think we're at the level of maturity that we want. I think a long-term project progression would be that we first get really comfortable with the models as they work more at a portfolio level. Quite honestly, I think with some of those models, that's the only level that they work at. The models don't agree very much, so there's a lot more work to do. We are still in the early stages of spending a lot of time understanding the individual models, and then figuring out how best to apply them.

Some models are better in certain parts of the country than others, for example, inland vs. on the coast. We'd like to see how they work together and determine what is the best blend in order to come up with something that we think is, I hate to say accurate, but more representative at the property level. When I get modeling output from any of these firms, I can look at where the water flows, and see how that model treats it, for instance. If a model makes it look like water is flowing uphill, well, that typically doesn't happen, right? It's not just about buying some models. We have to spend a lot of time building out the data and the capabilities to really understand the strengths and weaknesses.

Do you know how long assessing the models will take?

The U.S. government has a lot of this data, if they ever made some of it public to the extent possible, maybe that would make all models better tomorrow, but it's hard to really ballpark it. We have got to look at the gating criteria and say in order for us to use it for something, it would have to meet these following standards. For some of those usages we are years off.

Could your work be applied to loss mitigation?

I could see in the future that it could possibly be. Areas more likely to need forbearance are those more likely to have short term disasters. So I think that could be part of it. The analytics do help you understand what happens to borrowers after a disaster. You find that if a disaster hits an area that has insurance, the impact is not very big. If you look at the impact of Harvey, which was 80% or 90% outside FEMA special flood hazard areas, then you see a difference in performance. You see a higher delinquency likelihood and a higher default likelihood.

Does the range of models you’re examining extend to events outside of “traditional” disasters, long-term impacts and energy transition risks?

Energy transition is part of climate risk so we do look at some of the models there. We also look at our own internal analytics and say, okay, is there a scenario that we should run that if X and Y happened, what would happen to the performance of mortgages? Certainly, we look at both what's called the physical risk and the transition risk due to climate, because generally the quicker you address the climate risk with energy improvements, the less long-term physical risk there is. So we do look at how both of those things balance out.

I know you said a lot of your work is on long-term issues. Is there a short-term goal you have now?

Our goal is how can we really get a full understanding of what the risk is and feel comfortable with that? That's a clear goal.

We also have goals related to just raising awareness. We spend a lot of time testing out different messaging into markets, to see what resonates with borrowers and renters to see if they understand the risk and what keeps them from taking action. I see an awareness gap across the United States where people don't understand the level of risk they're facing, and once they understand their level of risk, they don't know what actions they could take.

We just got done with our latest flood survey. It looks at the awareness of flood risk across the United States, and I think we'll publish that later this year, but the findings are that the gaps are still huge. Most people don't understand what their flood risk is before they buy a property and generally we find that that skews lower on minority and low income areas. That's something we're actively working on and trying to set some sort of goals for, to get into the communities and really understand and refine the messaging to see that it actually starts taking hold.

How do you engage in community outreach?

We go into markets and try to get people to understand their flood risk. What we got out of that is we understood what messages resonate more with borrowers, and so this year, we're going to go through the same process again and further refine that message. Hopefully in a year or two, we'll be able to track whether we see the amount of people buying insurance go up. One thing we ask is do you know what flood zone you are in? Most people get that horribly wrong. What I'd like to see is that people really understand what risks they have, and make really thoughtful decisions about whether they should make their property more resilient or buy insurance.

One survey that we did on flood was a panel where I think they had like a half an hour discussion with folks. It was pretty in-depth. We may follow it up this year with also focus groups about which messages resonated and which didn't. Can we understand why? Other things we work on are Facebook and Instagram messages. We'll do Google searches to see what resonates and put banners up there saying things like, hey, did you know how many households are at risk of floods? Or what the average flood claims is, and to click on the question if they want to know more. Then we can see how many clicks we get. We show people what their flood risk is and we provide a lot of materials for them to educate themselves about taking action.

We haven't used TikTok yet. Maybe we should, but it depends on a lot of things including budget. The medium you used can be dramatically different in terms of the pull-through rates. One of our findings was that Facebook messages really resonated with seniors, who in some cases were one of the hardest groups to reach. In some places after disasters we've found what works best are radio ads. We try to make sure we're up high in Google searches for disaster recovery so people know who to call.

What would you say is your biggest challenge?

I think if you ask anybody in climate risk about that, they'll say data availability. The way I describe that is in terms of updates for the appraisal form. People never thought the elevation of the property was really important but for flood risk it is. Whether a property is built with a certain pitched roof is important for wildfires and wind. So the acquisition systems for data weren't built to capture a lot of the things that are important when it comes to sizing up climate risk. So it's going to take awhile to get not only a standard property data set that works for valuation and compliance, but also for what does climate risk look like?

Everybody in our bubble really understands climate risk. I think the challenge of the next few years is to make sure that every American understands it and I think disclosures are really important. I think getting properties more resilient to climate change is going to be probably one of the biggest challenges we face in the next 10 years because the building codes have not historically been up to standard. In the next 30 years, things are going to only get more harsh, so we've got to figure out a way of making properties more resilient.