Fast-forward to today, when the U.S. housing market is again having a historic run. On Tuesday we learned that U.S. home prices between March 2021 and March 2022 soared 20.6%. That’s the biggest 12-month uptick ever recorded, and well above the peak 14.5% year-over-year rate posted in the years leading up to 2008. The fastest-growing markets, like Boise and Austin, have even seen bigger hikes.

This housing boom has also once again seen many U.S. housing markets become detached from underlying economic fundamentals. At least that’s the finding from three separate proprietary housing assessments that Fortune has received over the past month from Moody’s Analytics, CoreLogic, and researchers at Florida Atlantic University. While many housing economists say it’s premature to call the pandemic housing boom a full-blown housing bubble, it’s certainly worth questioning how regional housing markets would fare if a housing slump did hit. The fact that spiking mortgage rates this spring are finally creating the conditions for cooling makes posing the question now even more important.

To answer that, Fortune teamed up with Home.LLC, a startup that provides down payment assistance to homebuyers in return for a share of any profits. Over a six-month period, data scientists at Home.LLC helped us build a forecast model to assess markets’ risk factors. In total, we looked at the 100 largest metropolitan statistical areas in the nation.

To calculate the risk assessment, we factored in 14 metrics. Those metrics include building permits, land use restrictions, delinquency rates, housing inventory, and average home tenure. (To see a full list of weights used in our risk assessment, go here.) Finally, we grouped regional housing markets into three tiers: low risk, moderate risk, and high risk. The housing markets labeled “high risk” would, by our estimation, fare the worst if a housing correction or crash materializes over the next few years.

Fortune and Home.LLC found that many of the housing markets with the highest risk levels are once again in many of the nation’s fastest-growing markets. In total, 11 regional housing markets received our "high risk" label. Those markets are Atlanta; Boise; Cape Coral, Fla.; Deltona, Fla.; Des Moines; Jacksonville; Lakeland, Fla.; Miami; North Port, Fla.; Orlando, and Winston Salem, N.C.

Of those 11 "high risk" housing markets, nine are located in the nation's booming Southeast corridor. Homebuilders in that region got pushed into overdrive over the past two years. As workers in high-cost Northern cities like New York, Boston, and Chicago realized during the pandemic they’d be able to do their job from home for the foreseeable future, many departed for the relatively low-cost Southeast. The weather certainly was a selling point too. A similar phenomenon happened in Boise as it became the top destination for professionals departing California over the past two years. But if working from home loses momentum, so could places like Boise and the Southeast corridor. Already, offices in the nation's largest cities, as measured by Kastle Systems, are seeing steady month-to-month increases in foot traffic. If the Federal Reserve’s inflation fight ultimately leads us into a recession, it could also see employers finally regain the economic leverage to force more workers back into the office.

The underlying reason that markets like Jacksonville, Atlanta, and Orlando are rated as "high risk" by Fortune and Home.LLC is those places have high levels of homebuilding, leaving them at a higher risk of oversupply if a housing correction comes. If you told homebuyers who have lost bid after bid in those low inventory markets that they’re at risk of oversupply, they’d probably tell you that you’re crazy. But so would have 2006 home shoppers in Phoenix and Las Vegas.

Of the 100 housing markets we looked at, we determined that 37 housing markets have a "low risk" profile. In other words, it's unlikely they would see a significant bust if a housing correction comes along. Those markets include places like Boston; Columbus; Denver; Grand Rapids, Mich.; Los Angeles; Portland, Ore.; Sacramento; San Diego; San Jose-Sunnyvale; Seattle; and Honolulu. The reason that markets like San Jose and Boston get a "low risk" label is that our model suggests the low levels of building in those places could buffer them from a steep housing crash. Limited supply coupled with favorable demographics (all those first-time millennial homebuyers) could prevent those places from entering into a free fall.

For a second opinion, Fortune reached out to Moody’s Analytics.

Moody’s Analytics chief economist Mark Zandi did his own analysis, looking at both overvaluation and how susceptible regional housing markets would be to a price drop. The finding? Among the nation's largest 392 housing markets, Moody’s Analytics finds 96% are "overvalued" relative to what incomes would historically support, while 149 of those housing markets are "overvalued" by more than 25%. Meanwhile, Moody's Analytics projects the following regional housing markets have the highest odds of a home price dip: Boise; Colorado Springs, Colo.; Las Vegas; Coeur d’Alene, Idaho; Tampa; Atlanta; Fort Collins, Colo; Sherman, Texas; Jacksonville; Idaho Falls, Idaho; and Lakeland, Fla.

Zandi thinks that price dip could be just around the corner. In fact, Zandi says soaring mortgage rates have already pushed the U.S. housing market into a "housing correction." By this time next year, Zandi expects year-over-year home price growth to drop to 0% and the nation's most "overvalued" housing markets to see home prices fall between 5% to 10%. If a recession comes, he says those price dips could quickly rise to 10% to 20%. Coming from a buttoned-down economist like Zandi, that’s quite the proclamation.

"In terms of home sales, they're falling sharply. Housing demand is coming down fast. Home price growth [will] go flat here pretty quickly. We will see [home] price declines in a significant number of markets," Zandi told Fortune. "Everything points to a rolling over of the housing market."