Why I Changed My Mind About Housing Bubbles

Revisiting an old theory in light of what the pandemic — and our own local data — taught us about prices, value, and the dollar itself.

By Tim Bray •  Broker / Owner  •  Seaport Real Estate Services

Years ago, I published a piece on this blog about housing bubbles, market crashes, and the idea of mean reversion — the belief that home prices, like an object thrown into the air, must eventually return to a long-term trend line. At the time, fresh off the 2008 crash, it felt like settled truth. Prices had soared, then fallen hard, right back toward the trend. The theory fit the evidence.

I no longer believe that theory holds the way I once described it. The years since the pandemic have run a real-world experiment on it, and the result was clear. So here is what I got wrong, what I now believe, and why it matters for anyone buying or selling a home today.

What I argued before

The original article rested on a single powerful idea: mean reversion . I argued that after a period of rapid appreciation, home prices would inevitably fall back to a trend line, and I leaned on a related assumption — that prices were tethered to local incomes and would have to revert toward what families could actually afford. The further prices ran ahead of incomes, the more “overpriced” the market supposedly was.

The logic was sound on its own terms. The problem was a hidden assumption underneath it — and a second one I didn’t examine closely enough.

What the pandemic disproved — and what it didn’t

Here is where I have to be careful, because the honest story is more interesting than “I was wrong.” Mean reversion is real . After the 2008 credit bubble, prices in our own market fell and then went sideways for nearly a decade. That is reversion, exactly as the original article described. I wasn’t wrong that markets can revert.

What I got wrong was assuming the post-2020 surge would behave the same way. It hasn’t. Prices stepped up to a dramatically higher level and have held — and, tellingly, they held even as mortgage rates climbed from roughly 3% to over 7%. A credit-driven bubble pops when rates rise. This didn’t. That tells you the post-2020 rise was driven by something other than cheap credit and speculation.

My trend line, and my income anchor, were both measured in dollars — and I quietly assumed the dollar was a fixed, stable yardstick. It isn’t. During the pandemic, the U.S. money supply grew by roughly 27% in 2020 and 2021 alone— the largest two-year expansion in American history, bigger than during World War II or the 2008 crisis. When you add that many new dollars to the system, each dollar buys less. Prices measured in those dollars step up and stay up.

Where I was right about income — and where I wasn’t

I want to be precise here, because it would be easy to overcorrect. I was not wrong to connect home prices to incomes. That relationship is real, and it explains the genuine pain families feel today: when prices race ahead of wages, housing becomes less affordable, full stop. That part stands.

What I got wrong was treating income as the gravitational center that prices must fall back toward. Price and income can decouple — and stay decoupled — because the dollar is shifting underneath both of them. Prices track the money supply; wages lag it. The gap between them is the squeeze. But that gap doesn’t necessarily close by prices falling back to incomes. Income is still half the story. It’s just the lagging half, not the anchor.

What this looks like in our own backyard

I don’t have to argue this in the abstract. Using our Market Pulse data, I pulled the median selling price for four southeastern Connecticut towns going back years — two on the water (Stonington and Waterford) and two inland (Ledyard and Colchester). If the post-2020 jump were just a rush for waterfront, the inland towns should have lagged. They didn’t. All four stepped up together.

Two things stand out. First, the step-up was regional , not coastal — inland Ledyard and Colchester rose as much as the waterfront towns, which points to a broad force lifting everything, not a location preference. Second, and even more telling: across all four towns the price per square foot roughly doubled from 2019 to 2025, even as the typical home that sold got smaller . People paid far more for less house. That’s not “nicer homes are selling.” That’s the same square foot repricing in a diluted dollar.

A fair caution: these are four towns in one corner of Connecticut, sharing many forces beyond the dollar. I offer them as a vivid local illustration of the national picture, not as proof of causation. But they line up with the national data unusually well.

The divide that actually matters: owners vs. renters

Money creation doesn’t reach everyone equally. People who already owned assets — homes, especially — saw those assets reprice upward. People holding their wealth in paychecks and savings fell behind, because wages and cash react last. The real dividing line isn’t rich versus poor. It’s owners versus renters— whether your wealth sits in an asset that floats up with the money supply, or in dollars that quietly lose value.

There’s a newer force hardening that divide, worth understanding without exaggerating. As high prices and rates pushed home purchases to a 30-year low, investors stepped into the gap: investors of all sizes bought roughly a third of single-family homes sold in mid-2025, the highest share in years. When a family is outbid by a buyer who never intends to live there, that family often moves from the owner column to the renter column — paying rent on the very asset class appreciating away from them.

But the popular version of this story is overstated, and honesty matters. The big Wall Street landlords everyone pictures own only about 3% of single-family rentals nationally; the very largest own around 2% of the market. The dominant players are small “mom and pop” investors. The institutional buyers cluster in a handful of Sun Belt metros — Atlanta, Charlotte, Tampa — where their local share runs high, while in most of the country, including Connecticut, their footprint is under 1% of the housing stock. So far, this is something we’re watching arrive elsewhere, not something reshaping our market. And in fairness, institutional capital isn’t purely a villain: it funds new build-to-rent supply and has been a net seller of homes for the last six quarters. The concern isn’t that investors exist — it’s that converting owned homes to rentals at scale can lock a rung of the wealth ladder away from the middle class.

What might happen next

The original article’s real sin wasn’t the theory — it was false confidence in a single prediction. I won’t repeat that mistake. The next several years could unfold along several very different paths, and no one knows which. Here are four worth holding in mind:

The wildcard: artificial intelligence

I’d be remiss to project a decade forward without naming the force that could bend every one of those paths: AI. And it genuinely cuts both ways.

On one side, a real AI productivity boom would make goods and services cheaper to produce — that’s disinflationary , pushing against the very dollar-dilution this piece describes. It’s possible the dollar’s value becomes a tug-of-war between money printing and AI-driven productivity rather than a one-way slide. On the other side, if AI suppresses wages across knowledge work, it sharpens the owner-versus-renter divide: people whose wealth is in labor get squeezed while those whose wealth is in assets capture the gains.

I’m not going to pretend to know how that resolves. If there’s one lesson from the history of transformative technologies, it’s that confident predictions about them age badly — in both directions. That uncertainty is precisely why I’m offering a range of outcomes rather than a forecast.

What I still believe — and what I now believe instead

Plenty from the original piece still stands: real estate moves in cycles; supply and demand still rule; speculative excess and over-leverage are still dangerous; and recent performance is still a poor crystal ball. None of this is a license to overpay. The discipline I preached then is, if anything, the part that aged best.

What’s changed is the frame. The question isn’t “are homes overpriced and due for a crash?” It’s “which side of the dollar do I want to stand on?”A fixed-rate mortgage is the rare tool that lets an ordinary family stand on the asset side: you lock in today’s price and repay the loan over decades in cheaper future dollars, while the home tends to rise with the broader money supply. It’s also how you opt out of paying rent to whoever ends up owning the asset. Sitting in cash, waiting for a reversion to an old trend, is itself a bet against the dollar — and history has rarely rewarded that bet.

This article revisits and corrects “Housing Bubbles & Market Crashes: The Holy Grail,” originally written by Barry Neilsen and modified by Tim Bray. Updated and corrected by Tim Bray. For general education only; not financial, tax, or investment advice.

Town figures from Seaport Market Pulse; national figures from the Federal Reserve (M2SL), S&P Case-Shiller (CSUSHPINSA), Brookings, the Urban Institute, and the GAO, rounded for clarity.