If you’re among the sundry folks encouraged that stocks just regained bull market status, that the Fed appears near the end of its tightening cycle, and that we’ve ducked a widely predicted recession so long that it may not be coming, here’s a data point to douse your cheer: Based on the traditional measures, the "sheetrock and framing" basics that normally determine its worth, the U.S. housing market is overvalued by 40%.
Shocking as it seems, that’s a view that should carry a lot of weight, because it comes from of one of America’s top real estate sages, Sean Dobson, founder and CEO of property powerhouse Amherst. For this writer, Dobson offers the best combination of on-the-ground proprietary data, gleaned largely from his own business, plus a mastery of sophisticated analytics, of anyone in the field. Amherst manages $16.9 billion in capital for investors, and ranks among America’s largest owners of single-family homes for rent, overseeing a portfolio of roughly 44,000 residences across 32 metros in 19 states. Dobson is a pioneer in deploying artificial intelligence to automate the appraisals and purchases of houses in those sweet-spot venues, an innovation I explore in my 2019 story, “Meet the A.I. Landlord That’s Building a Single Family Home Empire.”
Dobson also runs a big mortgage-backed securities trading operation, and MBS hedge funds that have regularly garnered spectacular returns. During the 2005 to 2007 housing bubble, Dobson took huge short positions in risky “Alt-A” home loans that, when the market imploded, generated $10 billion in gains for his investors.
Housing market predictions for 2023
It's crucial to note that Dobson’s take that prices are excessively elevated doesn’t mean they need to collapse going forward, as they did in the aftermath of the Great Financial Crisis. He cites a number of positive, countervailing forces that should greatly cushion the adjustment to come. As he told Fortune, we do know how that prices are highly inflated right now by historical standards—that’s a certainty. But though Dobson believes that plus factors, including a fall in rates and a continuation of relatively strong incomes, will prevent steep declines, those are still forecasts that may not materialize. “Is this another 2009 and 2010? We don’t think so, probably not even close,” says Dobson. “But when this price condition occurs, the world is a dangerous place.”
Amherst’s estimate of that 40% over-valuation doesn’t apply to the entire United States, but to the nearly three-dozen cities where Amherst purchases and leases single-family houses. Its portfolio features heavy concentration in Sunbelt cities that count among America’s fastest growing markets, including Tampa, Atlanta, Charlotte, Dallas and Phoenix, as well as affordable Midwest stalwarts, among them Louisville, Cincinnati and Kansas City. Even though Amherst doesn’t cover gateway metros such as New York, L.A. and San Francisco, for this reporter, its numbers are a pretty good proxy for the overall national market.
Amherst determines whether housing is under or over-priced by comparing two sets of data. The first comprises the actual, real-time prices across the Amherst markets. The second is the estimate of “fundamental” or “fair” value furnished by its proprietary model. The two most important metrics in that calculus are interest rates and personal incomes, though other factors, including demographics, also play a part. The difference between the two metrics shows the divide, as Dobson puts it, “Between what houses are selling for, and what our model says they should be selling for.”
The gaps are what he calls the “green icebergs” that rise above the zero line (overvalued) or sink below “water” (undervalued}. As the graph shows, a huge divergence opened starting in mid-2003 during the boom, peaking in mid-2006, when Amherst displays prices as too high by 35%. “That’s when we put on the big short on mortgages,” says Dobson.
After the crash, the market entered a kind of golden age of rationality and calm. Actual prices and fair value rose in a parallel, on a gradual, upward path, over the entire period from early 2009 to the start of 2021. “In that span, prices rose by about 75%,” says Dobson. “But that was perfectly explainable by changes in interest rates, income and demographics, all the things that go into our models. You were always at slightly plus or minus fair value,” as illustrated by by the two lines moving in harmony.
The boom starting in the opening months of 2021 drove prices skywards by 17% through the summer of 2022. “That’s what exotic monetary policy will do for you,” marvels Dobson. They’ve been flat, but not declining, since hitting that pinnacle, defying the spike in mortgage rates. But fair value, hammered by fast-rising home loan costs, went the other way fast, tumbling more than 20% from the start of 2021 through May of this year. The black and red lines heading opposite ways fast created a tall, sharp-peaked green iceberg looming at an “over-valuation” level of 38% to 39%. “That’s higher than at the 2007 summit,” says Dobson. “It’s remarkable that even though mortgage rates rose sharply, people still aren’t getting any discount on home purchases. The entry point for the average consumer buying a house isn’t very attractive.” The opening of that gigantic gulf led Amherst to substantially lower its monthly purchases of single-family homes for rent.
What's the current state of the housing market in the U.S.?
Given that his model puts today’s prices at almost 40% above housing’s fundamental value, Dobson’s astounded that we’ve seen few declines in his markets—keeping in mind that Amherst invests not in the likes of San Jose or L.A. that have fallen sharply, but in sunbelt cities that remain relatively affordable. “Likewise, every economist in the world would have told you that if you increase rates this quickly, prices are going to fall significantly. And they didn’t. So that’s the mystery. How long can that condition persist?”
Fortunately for homeowners, Dobson sees fundamentals swinging back in their favor. Hence, the huge fall the model found plausible, and that didn’t occur, doesn’t have to happen.
The first and most important likely tailwind is a decline in rates. “The model has fair value climbing back, mainly because the ‘yield curve’ is inverted,” he says. “That implies that mortgage rates are going to come down a lot.” For example, Treasurys maturing in 10 years carry a yield 3.7%, 1.7 points below the number for six-month bills. Since mortgages are typically priced at 1.2 points over the ten-year, that implies that home loan rates should drop in the next couple of years to around 5%, way below today’s figure of 6.9%. “People are putting hundreds of billions at risk betting that rates are going down and they’re probably right,” says Dobson. He adds that the decline in carrying costs should go a long way towards bridging the great divide between prices and fair value.
The second factor: Stable income growth and employment. “What’s flying in the face of everyone’s forecast is the strong labor market,” observes Dobson. “Consumers were already well funded because of the big stimulus. And now, the ‘voluntary quits’ number remains high because people are so confident they can get another job that pays more, they’ll leave their current position without another one lined up. And they often get about a 10% increase when they change.”
Lower rates, accompanied by rising middle class incomes should help stabilize home prices. But a recession that lowers both incomes and rates won’t do the trick. “You need rates to go down and incomes not to go down to cure this problem,” concludes Dobson. His “we escape the over-valuation problem” scenario hinges on avoiding a downturn that sends incomes crashing, the disaster movie that remains a possibility.
Housing market trends and inventory
Dobson also ran through Amherst proprietary stats on inventories, from Amherst’s tabulation of third party data as of April 2023. They display an astounding decline in homes-for-sale, and show the factors driving the phenomenon. In April of 2019, Amherst’s 30-odd markets offered 545,000 listings. By April of 2022, as the market still ran hot, the number had dropped by 42% to 318,000. The main reason: Houses were changing hands at a fast pace, with 78% selling within 90 days. In April of this year, the for-sale-sign count is still extremely low at 325,000 (41% below 2019), but this time the reason isn’t that homes are selling unusually fast, but that so few abodes are entering the market. The volume of fresh listings posted in the past 30 days was 30% lower in April than a year earlier, and 40% under April 2019.
The explanation, says Dobson, is basic: Families who benefited from ultra-low-cost mortgages over the past decade just won’t move. “When you give home buyers loans that are way below the fair market rate, from a supply and demand perspective, you’ve just ‘burned down’ that house. It’s gone from the market.” Bargain home loans have people trapped in their colonials and ranches, whether they want to be or not. “Houses are arguably much too expensive,” says Dobson. “That’s what you’d predict if you bring down the supply that much. The wonder is that people are buying them anyway,” though not nearly as fast as eighteen months ago.
How WFM has changed the housing market
To explain why prices remain so high in his markets, in defiance of his traditional model, Dobson considers a possibility he’d always rejected before: Maybe this time it’s different. “In the past when I’ve heard that maxim, I’ve always said, ‘It’s never different,’” he declares. “But now, the house is becoming the beneficiary of all the lost occupancy in the rest of real estate, notably offices. We still have the same number of people. They didn’t all go to Mars along with Elon Musk. They’re spending much more time at home. They typically used to spend 40 hours a week in their offices and now they could be spending 30 of those hours at home.”
A couple both working at home, sharing a one-bedroom apartment, may not work anymore, he adds. Now, they’ll pay a premium for extra space and a garden that they wouldn’t shoulder before the pandemic changed how people live and work. “We had people who moved from Manhattan to Richmond and I didn’t even know it!” recalls Dobson. “I’m on a Zoom call, and I see a pool in the background, and she says, ‘Yeah, we moved from a one-bedroom in Manhattan to a house on a golf course in Richmond.’ It was classic. She makes the same pay, but she’d exported an urban income to a suburban market. People used to live 30 minutes from their office, now they might live a three-hour flight from their office. It’s kind of wild, right?”
For Dobson, the new willingness to pay lot more than before to own a house than rent an apartment rent might explain why the chasm between actual prices and fair value remains so persistently wide. “For now, the work-from-home revolution has changed the demand equation for the asset,” he says. “The question is, ‘Is it permanently different this time?’”
Housing market projections through 2025
Dobson’s analysis, as shown on the price-vs-fair value graph, includes not just past and present data, but projections out to the fall of 2025. His analysis does not include the new paradigm, the this time it’s different possibility. He’s using the traditional model of fundamental value relying heavily on projections for rates and incomes. It shows prices falling in the 10% range over the next two-and-a-half years. That decline would eliminate most of the price-to-value gap. It’s hardly a rosy outlook, but it’s a relief when you consider that today’s green iceberg stands higher than the over-valuation mountain of 2007. He cautions that a large potential “error” around his forecasts means that, for example, if the market’s wrong about a coming fall in rates or we get a recession, the drop could be far worse. But on the opposite scale sits the possibility that housing has just experienced a big, during jump in value via a revolution in how people live and how they value livings spaces, that won’t go away.
For this hard numbers guy, the data savant who’s feasted from wagering against past fantasies about housing, the idea that it may really be different to this time, for once, doesn’t seem like just another pipe dream.
NOTE: HPI Growth utilizes Amherst’s Housing Price Index (“HPI”), which tracks U.S. level home price changes based on the weighted average of corresponding state-level home price indexes. The index is based on the Case Shiller repeat-sales methodology. Unlike other HPA indices, the Amherst HPA does not include foreclosures, shortsales, bank repossessions, and REO resales. The use of Multiple Listing Service (MLS) to supplement CoreLogic off-market data enables Amherst to offer a timelier look at shifts in the housing market than other indices.
