Supply and Demand Signals for Real Estate in 2025

The housing market confuses people because it moves slowly and then suddenly. You can watch every indicator point toward distress for eighteen months and prices still don't budge. Then something shifts and it moves fast. That lag is where the opportunity is, and it's also where most people get burned by assuming the trend will continue when the underlying dynamics have already changed.

The way I think about real estate is that it's supply and demand with a very long lag built in on both sides. Supply takes years to respond because permitting, financing, and construction all take time. Demand shifts slowly because people don't move their families based on monthly rate fluctuations. When those two slow-moving forces get out of sync, prices do the adjusting. Right now they're out of sync in interesting ways depending on where you're looking.

The Rate Lock-In Effect Is the Single Biggest Supply Constraint

Just under half of all outstanding US mortgages still carry a rate below 4%. That's 49.9% as of Q1 2026, per FHFA's National Mortgage Database, the first time the share has been under half since 2020. At the peak in early 2022 it was above 65%. With 30-year fixed rates hovering around 6.5%, someone who bought in 2020 or 2021 faces a real economic decision: sell, move, and take on a mortgage at roughly double their current rate, or stay put. Most are still staying put.

This is the rate lock-in effect, and it's been the dominant force compressing existing home inventory since 2022. It is melting, though. Existing-home months of supply hit 4.6 in June 2026, with a median price of $440,600. Six months is roughly a balanced market, so 4.6 is still a seller's market, but it is a long way from the sub-three readings of 2022 and 2023.

Here's the part people get wrong. The lock-in effect unwinds gradually, not suddenly. Below-4% mortgages have been shedding share for four straight years without anything resembling a wave. As rates ease, the gap between outstanding and current rates narrows and more homeowners can make the economic case to sell. Life events force moves regardless of rate math. But nobody voluntarily hands back a 3% mortgage. That supply comes back over years, not quarters, and any thesis that needs it to arrive faster is a bet, not a plan.

New Construction: The Regional Picture Is What Matters

National housing start numbers obscure enormous regional variation, and the national number is mostly useless for making actual investment decisions. What's interesting is the spread.

Texas, Florida, and most of the Mountain West have added housing supply aggressively since 2020. Lighter permitting regimes, lower labor costs, strong migration inflows. Austin added more housing units per capita from 2020 to 2024 than almost any other major metro. Prices there have corrected more than 15% from 2022 peaks. That's supply working. When you build enough, prices respond.

California, New York, and the Northeast have added far less supply relative to demand. Restrictive zoning, NIMBY opposition to density, multi-year entitlement processes, and high construction costs. San Francisco has seen net population decline and still maintains housing affordability ratios near their worst levels in decades. That's what happens when supply genuinely can't respond to demand. The price signal just stays permanently distorted.

The regions that matter for long-term appreciation are the ones where supply can't catch demand. Not because it won't, but because geography, politics, and permitting make it structurally difficult. Coastal California, parts of the Northeast, specific mountain resort markets. In those places, the supply constraint is durable. In markets where supply can respond, you're buying into a more competitive dynamic where any demand spike gets met with new inventory.

Zoning reform is gaining real momentum in a way that felt unlikely five years ago. California's ADU laws have added meaningful supply. Oregon ended single-family-only zoning statewide back in 2019. Montana passed a sweeping reform package in 2023, got it enjoined by a district court almost immediately, and then had it upheld by the state Supreme Court in 2024. That sequence is the whole story in miniature: the policy passes, the local opposition sues, and the units show up years later if at all. Which is exactly why it's a durable tailwind and a terrible trade. You can't time it.

Remote Work Normalization and Which Markets It Makes Structurally Interesting

The remote work trade of 2020 and 2021 was mostly noise. People moved to vacation towns, bid up prices, and a lot of that reversed when companies started requiring office attendance again. What's left after the reversal is more interesting.

Remote and hybrid work has settled at roughly 25-30% of working days for knowledge workers. That's not enough to make a full relocation to a mountain town rational for most people. But it's enough to make a move to a second-tier city rational if that city has the infrastructure, amenities, and job market to support it. Raleigh, Nashville, Salt Lake City, Austin (before the inventory surge), and parts of the Mountain West benefited from this dynamic in a way that looks durable.

The markets that I think are structurally interesting for the next decade are the ones where remote work normalization created genuine demand migration, and where supply hasn't yet fully caught up. You're looking for cities where in-migration is still positive, where local job markets are diversifying beyond any single employer, and where permitting constraints are real but not California-level.

Commercial Real Estate: Two Very Different Stories

Talking about commercial real estate as a single category is almost meaningless right now. Office and industrial are in completely different worlds.

Office is in a structural reset that isn't going to resolve the way prior downturns did. Hybrid work landing at roughly a quarter to a third of working days is real, permanent demand destruction. Sublease availability is elevated and distressed sales have accelerated, with more than 200 distressed office assets trading in 2025 alone. The individual prints are brutal: downtown towers changing hands at 70% or more below what they last sold for. This isn't a cycle. It's a repricing. Some of those buildings get converted to residential or life science use. Many won't, because the floorplate and plumbing math on conversion is far harder than the headlines suggest.

Industrial and logistics is a different story, though not the uncomplicated one people still tell. E-commerce fulfillment, last-mile delivery, and reshoring are genuine structural tailwinds, but the sector overbuilt into them: national industrial vacancy has climbed to around 7.5% and rent growth has stalled. Anyone still pitching industrial as a can't-lose vacancy-at-record-lows trade is quoting 2022. What's actually strong is the data center corner of it, where national vacancy sits near a record low of about 2% and the AI buildout has made power-adjacent land worth more than whatever building sits on it. Prologis, the biggest logistics landlord in the world, is committing roughly $8 billion to build 20 data centers, which tells you where it thinks the demand is (our AI infrastructure roadmap walks through where that capex is actually going).

Multifamily is more nuanced. Sun Belt markets that over-built in 2021-2023 are seeing rent growth stall or reverse as new supply comes online. Coastal markets with persistent supply constraints continue to see rent growth. The "rent always goes up" narrative from 2021 was correct on average and wrong in the specific markets that over-built. That's always how it goes.

How to Think About Positioning

For direct real estate investors, cap rates in some secondary markets have finally improved enough to make the math work. Markets where stabilized asset cap rates are 150 or more basis points above the 10-year Treasury offer a reasonable entry point. Markets where you're buying at a 4.5% cap with a 7% cost of debt need either a genuine value-add thesis or a strong conviction that rates come down meaningfully. That conviction needs to be explicit, not implicit.

For REIT investors, data centers are the cleanest structural story and industrial is a more crowded one than it was. Office REITs are distressed and cheap for a reason. The discount to NAV is real, but so is the secular headwind. Distressed doesn't mean it's time to buy.

The stress test that matters most right now: what does your debt structure look like if rates sit around 6.5% for three more years? The investors who over-levered in 2021 on short-duration floating-rate debt are working through that pain now. The ones who locked in long-term fixed debt are fine. Debt structure is the risk that matters most in this environment, more than location, more than asset class. That's the cyclical noise to cut through. The durable appreciation is in supply-constrained markets with genuine demographic demand. Everything else is timing the rate cycle.