Will 2023 Be the Year?

Caution Trail 2023-1

Our 2023 outlook is highlighted with two thematic research summaries:

Impact of sharply higher mortgage rates on expected home price appreciation and our roadmap for the risky home market cycle ahead
Deep dive into expected foreclosures in Lee County following the devastating impacts of Hurricane Ian

The DeltaTerra Journey: 2022 Review

DeltaTerra ended 2022 with exciting potential partnerships that we hope to cement and announce in the very near term. Looking back on the year, I am filled with gratitude for:

1. You - our clients, partners, and friends.
  Your time, curiosity, interest, and encouragement have made all the difference.
2. Our incredible team
  I'm humbled and deeply grateful that the DeltaTerra team has remained steadfast. Our highly skilled team of portfolio managers and engineers have chosen to devote their brilliance to our cause of bringing more climate risk understanding to the financial markets. Together, we are continuing to build and launch game-changing products. Some key achievements include:
  - Launch of our Klima Credit Risk Transfer (CRT) subscription service that provides bond buyers with workflow-relevant assessments of climate’s contribution to credit risk for every outstanding CRT security
  - Launch of our Klima Maps product that offers a visual representation of physical climate risk, market exposure, and asset depreciation risk estimates at the census tract or county level for every US home market – and the mortgage securities most exposed to the scenarios playing out in any selected group of metropolitan areas
  - Beta release of our Klima Commercial Mortgage-Backed Securities (CMBS) product supported by research into commercial real estate (CRE) market exposure to physical risk, underinsurance, and scenario depreciation risk for five property types in every major US metro area
3. Our investors
  We would not be here today without the confidence and support of our incredible investors and for that we are deeply grateful. Our longevity has allowed us to witness the changing of the tide in regard to public and investor opinion on climate change from distant possibility to a confirmed reality.
4. Our luck with timing
  Gone are the days when newscasts cover the debate on the existence and drivers of climate change. Now, public discourse centers on how to deal with the many hazards that are endangering communities around the world. Never has the time been riper for our solutions and we have been seeing that play out in our partnerships and sales efforts. We are delivering solutions that key early adopters have been searching for and many others are beginning to realize they need.

Risky Home Cycle Roadmap

Many investors accept the basic premise of climate mispricing in property values but struggle to incorporate the theme into investment strategies because the timing of the rationalization is so difficult to predict. We certainly don’t have a crystal ball into this issue, but we do have a roadmap and closely track the highest frequency market data to locate our position on that map and reevaluate the path forward as necessary. Below, we provide a description of this roadmap, including the rationale behind our dire predictions for risky home markets and our relatively constructive prediction for positive home price appreciation across the overall US in 2023.

Very low mortgage rates and work-from-home pressures during the COVID pandemic led to strong demand for risky homes. If you are going to spend a lot of time at home, you might as well have a pleasant view of the water or forest. In 2021, however, insurance prices began increasing dramatically for exposed properties just as mortgage rates began to rise. The affordability-led slowdown was significantly more pronounced for risky homes, as illustrated in the chart below of year-over-year sales growth for risky homes vs. safe homes. Data sources include property-level transaction data obtained from LightBox and physical climate risk data obtained from the First Street Foundation.

YoY Sales Growth

We currently include three types of climate repricing risks in our definition of risky properties, all driven by current structural transitions happening in hazard insurance markets:

Below is the year-over-year sales growth rate differential for homes with high wildfire risk (defined as having a FireFactor score from the First Street Foundation of seven or above) vs. those with low wildfire risk. High wildfire risk properties have been underperforming since January 2020, just after the expiration of the first emergency insurance non-renewal bans enacted by the California governor in 2019.

The next chart illustrates the demand growth differential for properties in FEMA-defined Special Flood Hazard Areas (SFHAs) vs. properties outside of the flood zone. Sales in flood zones boomed during the pandemic but then turned negative starting in July 2021 just a few months before FEMA’s Risk Rating 2.0 initiative kicked off to reprice National Flood Insurance Plan (NFIP) premium rates. This coverage is mandatory in SFHAs for a homeowner that takes out a mortgage from a federally regulated lender.

The third subsector of risky homes we are watching are homes outside of flood zones that would be good candidates for inclusion in new mandatory insurance requirements with modernized risk designations. Demand growth for these properties has stayed on par with safer homes outside of the flood zone, even after the Risk Rating 2.0 price hikes. Declining NFIP take-up rates suggest that buyers of high-risk properties outside the flood zones (where NFIP insurance is optional) are simply choosing to go uninsured rather than pay the higher premiums.

However, a new Florida law requires all Florida homeowners (in or out of the flood zone) insured by Citizens Insurance, the government backstop system, to purchase separate flood insurance. This will translate to a large increase in the number of high-risk homeowners outside of SFHAs for whom credit availability is linked to flood insurance, so we expect this chart to turn negative in the near future.

This demand underperformance in risky property markets is occurring in the midst of one of the most rapid declines in home affordability in history. Mortgage rates doubled in 2022 as prices continued to climb, sending the NAR affordability index below 100. A sub 100 reading on this index is meant to indicate that the median earning household can no longer afford mortgage payments at current interest rates. This has led to steep declines in home sales as illustrated in the next chart.

The downturn in the pace of home sales in 2022 was akin to the 2006 downturn that was precipitated by a much more modest increase in mortgage rates (but with growing unease about a credit bubble fueled by a subprime mortgage market that doesn’t exist today). In our Klima Base scenario, demand continues to weaken through 2025 as it did during the 2007-2009 period when concerns about declining prices discouraged buyers and lenders. Our forecast scenarios are also characterized by outsized weakness for risky homes that are also contending with sharp increases in insurance premiums.

While the current pace of demand destruction is worse than we saw at this point in the prior cycle, supply is a very different story. At the end of 2006 as the previous correction began, there were more than 3.5 million homes for sale. At the end of 2022 there were less than 1.5 million homes for sale. The building industry was badly hurt in the GFC and hasn’t recovered fast enough to meet demand as the rest of the economy recovered. This problem was exacerbated by the pandemic leading to the tightest market in documented history.

In our Klima Base scenario, supply ultimately outstrips demand in risky home markets leading to 2007-like fundamental price pressures but only in those risky markets. Even assuming continued demand pressure that follows the path of the last cycle when lenders vigorously tightened credit standards, supply barely gets back to normal equilibrium levels in safer markets. The chart below shows our forecasts for “months supply” or the number of months it would take to clear out the existing inventory at the current pace of sales.

Home price appreciation has a strong relationship with months supply as shown in the chart below of monthly observations of year-over-year home price appreciation vs. months supply six months prior. The green bar shows the most recent months supply reading for the US of 4.1. The market has rarely been this tight in history and negative appreciation has always been led by months supply reaching at least 5.

Applying the historic relationship between months supply and HPA to the high-risk and low-risk scenarios described above yields the below HPA projections. While we think asset values could start declining in 2023 for the 20% of homes exposed to structural changes related to climate risk, we think it is unlikely that prices will decline for the overall market this year.

Foreclosure Prediction for Lee County Following Hurricane Ian: ~4.5K Households at Risk

We used numerous data sources and some new climate-linked econometric models to estimate potential home foreclosures in Lee County in the wake of the catastrophic Hurricane Ian. Our mid case estimate puts 4.5 thousand households in the county at risk of foreclosure. This wasn’t exactly “six sigma” statistical work and should be treated with a high degree of uncertainty, but the confluence of concentrated damages and asset market headwinds could lead to a much more negative result than what we have seen following other recent catastrophes.

Based on early estimates of Ian insurance claims paid (from Verisk PCS), we model a ~6.6% distress rate for Lee County mortgage borrowers. We estimate that 129K of 242K Lee County homeowners have mortgages so this translates to ~9K households who will have some problems making mortgage payments in the six months following the storm. Of these troubled borrowers, some will find their way back to financial stability and others will face a foreclosure or short sale. Due to a confluence of many headwinds for Lee County real estate values, we expect a troubling outcome in which more than half of the distressed households lose their homes through foreclosure.

Below we provide a detailed description of our estimation framework broken out into deep dives into the two most relevant dynamics:

Relationship between insurance claims and the percent of mortgage borrowers that become distressed within six months of an event (defined as either reported delinquency or having to take advantage of disaster assistance offered by the agencies).
The extent to which the next two years of home price appreciation determines how many of those troubled borrowers are going to be able to get back on their feet and avoid a foreclosure (the “cure rate”).

Damage ⇒ Distressed Borrower %

We examined data on insurance claims payments and Freddie Mac loan pool performance for 39 large incidents since 2009 (defined as a major metro area getting hit by a hurricane that resulted in residential insurance claims amounting to more than 1% of the outstanding single-family home building value in a region).

We found a very strong relationship between the claims as a percent of stock measurement and the borrower distress rate that followed. Other factors that we found to be statistically significant in modeling borrower distress rates were the credit quality of the mortgage pool being examined (defined by average FICO score of the pool) and the region’s flood risk (defined by an average annual loss rate from First Street Foundation data).

Initial estimates of claims paid for Ian damages in Lee were 7.4% of the stock, a painful event by historic standards (in the top 15% of our “large incident” sample). Flood risk is very high in Lee at 1.1% AAL rate, and only 7 of our 39 incidents were in regions with higher flood risk. However, borrower credit quality is very high relative to the rest of our sample with current outstanding Freddie Mac loans in Lee having an average FICO of 751. Only three incidents involved higher credit quality pools so the net effect on the model is a lower expected distress rate per claims level in Lee than we would expect before considering flood risk and borrower credit quality.

The 6.6% distress rate is the intersection of the quality and flood risk-adjusted relationship (red line) and the realized 7.4% claims to stock ratio (yellow vertical line).

2Yr Home Price Appreciation (HPA) ⇒ Distressed Borrower Cure Rate

The degree to which distressed borrowers are able to find their way back to financial stability is highly dependent on market conditions. Specifically, if home prices are increasing in a region (despite the disruptions from the storm), borrowers are much more likely to make repairs and resume mortgage payments than if home prices are falling.

There are a lot of moving parts that contribute to the cure rate so we didn’t try to come up with an exact estimate for this exercise, but the historic relationship between asset market strength and cure rates is clear in the below visualization. We expect to see sharply declining home prices in Lee over the next two years as property markets price in much higher flood insurance costs. Our Klima model estimates market overvaluation of approximately 30% due to overly benign cost expectations for flood protection. If this overvaluation corrects over the next two years, we think it is reasonable to expect a sub-50% cure rate for impacted borrowers.

While this isn’t very scientific, the conclusion is easily visualized by extrapolating the below chart to some unplotted -30% point along the X axis. We highlight the experience of distressed borrowers in New Orleans after Katrina in 2005, where 91% cured as home prices increased by 16.5% in the region over the next two years vs. Gustav in 2008, where only 42% of distressed borrowers in the same region were able to cure as home prices declined by 4.4% in the GFC correction.

Home Prices Going Forward

The future path of home prices in Lee is of course uncertain, but the confluence of macro pressures from sharply higher mortgage rates and micro pressures from Lee’s flood insurance problems (which will be greatly exacerbated by the new Citizen’s requirement for flood coverage regardless of flood zone), is setting Lee County up for a very sharp correction.

After extremely high appreciation rates during the pandemic bump (monetary policy and WFH-driven), prices peaked in August of 2022 and have now started to turn.

While appreciation in Lee County has been very strong (Lee has historically been a “high beta” market), underlying supply and demand trends have been challenging, presumably since the implementation of Risk Rating 2.0 began to weigh on the metro area in 2021. Here is a chart comparing months supply (how many months it would take to clear out current homes for sale at the current pace of sales) in Lee vs. the whole US. From this perspective, Lee is in trouble.

Our Klima model suggests that the home market in Lee County is overvalued by nearly 30% vs. where it would be if property buyers started pricing in rational insurance cost hikes in the future. The current fundamental trends suggest that this is a very reasonable base case forecast in the ongoing correction. This last chart shows the relationship between months supply and six months forward HPA in Lee since 2018 when the inventory data from Zillow becomes available. Punch line – look out below:

We greatly look forward to working with our readers to better understand these important dynamics in markets and our economic systems. Whether you are a policymaker, investor, lender, or operator, we would love to hear your thoughts and feedback!

All the best,

Dave & the DeltaTerra Team

Published on Jan. 19, 2023.