How to Tell If a Housing Market Is Overvalued (5 Public-Data Signals)

"Overvalued" gets pinned on any market with a big price tag. That is the wrong test. A $600k market can be fairly valued and a $200k market badly overvalued. What matters is the gap between price and the fundamentals underneath it.

The good news is that the gap is measurable, and the inputs are public. Five signals do the work, and each maps to a number you can pull yourself: income, mortgage rates, inventory, sale prices, and a market's own price history. No black box, no vibes.

A market is overvalued when prices have detached from the things that actually pay for housing: local incomes, rents, the balance of supply and demand, and the market's own price history. Each signal below tests one of those anchors.

One stretched reading is a yellow flag with a dozen innocent explanations. Several stretched readings at once, all pointing the same way, are the real thing. Hold that distinction. It is the difference between a hot market and an overvalued one.

Prices do not float away from fundamentals at random. A few familiar forces push them there, and naming the force tells you how fragile the gap is.

The first two are the most reversible. A market floating on cheap credit or hot money can deflate fast when the rate cycle or the sentiment turns. Supply-constrained appreciation is stickier, which is why the signals below always get read against a market's own history, never a single national rule.

Each signal isolates one fundamental and asks whether price has run ahead of it. None of them requires anything you cannot pull from public data. Read them in order, then read them together.

One thing to hold as you go: the five do not all fire at the same time. The valuation ratio and the payment burden are leading. They tell you the standing risk a market carries before anything moves. Price-versus-trend is context.

Supply and momentum are confirming. They only flip once a top is already forming. A leading signal warns you; a confirming signal dates the turn.

The useful consequence: the two signals that buy you time are also the two you can check without watching a market week to week. Both come from figures published on a fixed schedule. Here is where all five come from.

Five signals, five public sources, no proprietary feed anywhere on the list. Pulling any one of these numbers is easy. Keeping all five current, for every market you are weighing, is the part that quietly stops happening.

Price-to-income is the median sale price divided by median household income. Both are public: the Census Bureau's American Community Survey for income, and local sales records or a public price index for the price. The long-run "balanced" range NAHB and CoreLogic tracked for decades was about 3 to 4. The typical U.S. metro has since drifted above it, closer to 4 to 5 today, so the whole country reads a notch more stretched than its own history.

When a single market pushes past 5, and well beyond its own past, prices have climbed faster than local paychecks can follow. That is the clearest single sign that value has run ahead of fundamentals.

Two honest caveats keep this from being a blunt instrument. First, it is a ratio of medians. The median buyer is not the median household, since buyers skew higher-income and renters pull the income figure down, so the raw level reads a touch high everywhere. That is exactly why the trend matters more than the number.

Second, for an investor there is a sharper cousin: price-to-rent. Dividing price by annual market rent compares the same asset on both sides, with no median mismatch, and a yield that compresses below your cost of financing is overvaluation you can feel in the cash flow. Read both; they answer the same question from two angles.

Every bar sits under the rate line, Pittsburgh included. At 6.6% financing no market on that list carries itself on rent alone, so the read is the size of the shortfall rather than its existence: 2.0 points in Pittsburgh against 4.5 in San Jose. That is more than twice as much of the price left unexplained by what the asset earns. Price-to-income tells a version of the same story, on a longer timeline.

The pitfall is treating one national number as gospel. Expensive coastal metros carry structurally higher ratios because land is scarce and incomes are high, and they can sit above 5, or above 8, for years without crashing. A market at 8x its income is not news; that is the coastal premium. A market that climbed from its own 3.5 to 5 in three years is the news. So read price-to-income against the market's own five-to-ten-year trend, not against Topeka: the move off the market's own baseline is the signal, the raw level is not.

Austin is the pin worth studying. At 4.3x it sits exactly on the median of all 980 metros, which is where a market lands after it corrects, and Austin ran this whole cluster at its peak. Boise at 6.0x and San Jose at 9.9x are out where only a few dozen metros live. Being in the tail is a fact about the level, and Austin is the reminder that the level moves.

Price-to-income ignores interest rates. This signal does not. It asks what share of the typical local income the monthly payment on a median-priced home would consume, using the prevailing 30-year mortgage rate, which the Federal Reserve publishes weekly. The long-standing affordability line is 30 percent of income, the same threshold HUD uses to define a cost burden. When the payment on a median home blows well past 30 percent of median income, the market is leaning on cheap credit or outside money rather than local demand.

Here is what makes this signal sharp: it can flip a market into overvalued territory with no change in price at all. When rates rise, the same home costs more every month, so affordability cracks while the sticker price holds. A market that looked fine on price-to-income can fail this test the moment the rate cycle turns.

Follow the green line past its 2022 spike and something useful happens. It crosses back under 30% in 2024 and keeps easing, while the rate stays between 6 and 7 the entire time. The relief came from the other side of the ratio: prices are 21.8% below their peak, and incomes have kept rising.

Cheap money never came back.

Months of supply is how long it would take to sell every home currently listed at the current sales pace, and it comes straight from public listing data. Four to six months is balanced. Under two months is a tight market. Tight supply is the mechanism that lets prices overshoot, because when there is nothing to buy, buyers bid each other up. Read it carefully all the same, since tight does not always mean hot.

Much of the 2022–2024 supply crunch was a lock-in effect: owners sitting on 3% mortgages would not sell into 7% rates, so supply stayed thin even where buyer demand had cooled. Persistent sub-two-month supply with prices still climbing is the pressure that pushes the first two signals into the red; sub-two-month supply with prices flat is often just frozen sellers.

Do not read tight inventory as a verdict on its own. A growing city with genuine demand can run tight for years and stay fairly valued, because incomes and rents rise alongside price. Inventory earns its place by explaining why the affordability signals are stretched, not by flashing on its own.

Two numbers catch a market topping out, and both are public. The sale-to-list ratio tells you whether homes still sell at or above asking; once it slips below 100 percent, buyers have stopped chasing. The share of listings with price drops tells you the same story from the seller's side. High prices paired with a falling sale-to-list ratio and a rising share of price cuts is the signature of a market that ran hot and is now turning.

One soft month is not a turn. Housing has a strong seasonal rhythm, and almost every market cools into the fall. Read the trend over several months, and against the same months a year earlier, before you call momentum broken. A genuine top shows up as a steady slide, not a single weak print.

The last test is the market against itself. Compare today's price to its five-year trend and to its all-time peak, using a consistent public price index such as the FHFA House Price Index. A market tracking its own trend is digesting growth at a healthy pace. A market sitting far above trend, or pinned near an all-time high while the other signals flash, has stretched.

This is the context check that keeps you from calling every expensive market a bubble. Price far above trend is only a warning when the fundamentals are not there to justify it. Pair this signal with the first two: above-trend prices plus stretched affordability is the combination that matters. Above-trend prices on rising incomes is just a market that grew up.

Here is the whole framework on one screen. The bands are rules of thumb, not bright lines, and the real read is always the cluster and the trend, not a single cell.

One row carries more weight than the rest when you are early. The payment-to-income row can walk a market from the left column to the right without a single price changing, because the rate does the moving. Every other row needs the price to move first, which is why they arrive late.

No single signal is a verdict. Each one has an innocent explanation: a coastal premium, a seasonal dip, a city growing into its prices. Any one of them, on its own, is noise as often as signal.

The read is the cluster. When three or four point the same way at once — price stretched, payments stretched, supply tight, momentum cooling — the innocent explanations run out. That configuration is what turns "expensive" into "overvalued."

Overvalued is a caution flag, not a stop sign. Plenty of money has been made in stretched markets and lost in cheap ones. The point of the signals is to make you price the risk instead of ignoring it.

So when the cluster lights up, do three things. Widen your assumptions, and underwrite flat or falling appreciation instead of a continuation of the run. Lean on the parts of the deal that survive a downturn: a strong rental yield and a long holding horizon. And let it sharpen your patience, because an overvalued market often hands you a better entry six months later, once momentum cools and the price cuts from Signal 4 do their work.

Here is the honest friction. Doing this right means pulling income from the Census, the mortgage rate from the Fed, inventory and sale-to-list from listing data, and price-versus-trend from a house-price index, then normalizing all of it, for every market you are weighing, every month. It is entirely doable. It is also a spreadsheet you will quietly stop updating.

Run the five signals across every U.S. metro, though, and the map of risk falls right out. Here are the most stretched large markets at the moment, the same framework applied everywhere at once:

That assembly is what LocalAlpha does for you. It pulls all five signals from the same public sources, scores every U.S. market on them, and rolls the affordability, inventory, and momentum reads into one number with the breakdown underneath. You read one score instead of maintaining five feeds. Every source and formula is published, for the same reason this article exists: no black box.

Here is where that cluster lands in the product. The Alpha Score is a general read on a market. It blends affordability, price momentum, supply, the local economy, and rental yield into one number, to surface markets worth a look. Overvaluation shows up in one of those bars in particular. Pull up Boise, and watch which bar is on the floor:

Boise is a clean case. Its price-to-income climbed from the low 4s into the 6s in a few years as out-of-state money poured in. That is a market climbing its own ladder, which is a different condition from a market that has always been expensive. The affordability bar is on the floor, dragging an otherwise-middling score down: that is where overvaluation lives.

But notice what the headline bars don't show. The general score won't surface the whole cluster on its own. Tight supply, one of the warning signs, actually reads as healthy here, because scarce inventory is good for an investment even as it props up an overvalued price. That is the tell that a general score answers a general question. To see overvaluation you read the affordability bar and the five signals together — the scorecard above, every value checkable against the sources listed at the foot of this piece. No black box; just the right lens for the question.

Now watch the other side of the same ruler. Austin ran this exact cluster at its 2022 peak: stretched ratios, tight supply, prices at an all-time high. Then it corrected more than 20%, supply ballooned back past five months, and the payment burden eased below the line. Today Austin reads balanced, not overvalued. The tell between still overvalued and already corrected is the supply bar: tight with prices near peak means the air is still in; loosening with prices off their high means it already came out. The 60-second read walks the same bars for any metro, and the most underrated markets sit at the far end, where price still trails the fundamentals.

Four years on, Austin still sits 21.8% below that May-2022 peak and reads as an ordinary market on all five signals. Note how long the far end of the ruler takes. The correction ran for years, and what it produced is a market that is simply typical now, which is the honest answer to what "overvalued" costs you: usually time rather than a crash.

Overvalued is not a feeling about high prices. It is a gap you can measure on public data, signal by signal. Run the five before you act on a market, not after.

P.S. The fastest version of this: pull up the market you are eyeing and the one you already own, and read the affordability bar on each. If the gap surprises you, the signals are doing their job.

Methodology: thresholds are standard conventions rather than our invention — NAHB/CoreLogic price-to-income, the HUD 30% cost-burden line, and the four-to-six-month balanced-supply rule. Figures are per Redfin (prices, supply, sale-to-list), Census ACS (income, rent), FRED (the 30-year rate) and FHFA (the house-price index), assembled per market by LocalAlpha; every formula and source is on the methodology page. Charts are snapshots as of the date each one states, not live reads.