A 19% jump in housing starts is the kind of figure that gets a one-line celebratory headline and no second look. In June it deserves the second look, because the number is technically accurate and substantively misleading, and the gap between those two things is the whole story.
Here are the facts. Privately owned housing starts rose 19% in June to a seasonally adjusted annual rate of 1.427 million, up from a revised 1.199 million in May and 3.5% above a year earlier. That is real. But look at what drove it. Multifamily starts, apartments and condos, soared 76% in a single month to a 513,000 pace. Single-family starts, the houses most Americans actually buy, slipped 0.2% to 895,000, a third consecutive monthly decline.
So the "surge" is almost entirely one volatile category bouncing back, while the core of the housing market kept shrinking. The deputy chief economist at First American said it plainly: the headline overstates the strength in homebuilding, reflecting a temporary surge in multifamily rather than broad recovery.
The dead-cat bounce nobody named
The 76% multifamily jump sounds spectacular until you see what it was jumping from.
The month before, May, was a disaster. Total starts had fallen to a six-year low, and multifamily specifically had plunged about 41% to its lowest level since late 2024. So June's 76% multifamily leap is not a new boom. It is a rebound off a collapse, the arithmetic of recovering from an unusually terrible month.
This is the trap in month-over-month percentages on volatile data. The Census Bureau says as much in its own fine print, noting it takes six months to establish an underlying trend for total starts, and that the 19% jump carries a margin of error of nearly 16 points. One bouncy month is noise, not signal.
The clearer read comes from the three-month averages, which smooth out the whipsaw. On that basis single-family starts are drifting down to around a 902,000 pace, while multifamily has trended up to roughly 445,000. Two different stories, and the headline blended them into one falsely cheerful average.
The number that actually predicts the future went the wrong way
Here is the part that should override the optimism entirely, and it is the reason to distrust the surge. Starts tell you what builders broke ground on this month. Permits tell you what they plan to build next, and permits fell.
Building permits declined 3% in June to a 1.367 million pace, and are down 2.3% from a year ago. Single-family permits, the forward signal for the market that matters most to ordinary buyers, dropped 2.4% to 871,000 and are running near the lowest levels of this cycle. One analysis put it bluntly: permits remain near cycle lows, and June's gain was multifamily volatility, not a clear single-family rebound.
Sit with the divergence. Starts jumped 19% while permits fell 3%. Builders do not pull permits for homes they do not intend to build, so a falling permit count is builders quietly signaling they expect to build less, not more.
Why builders are pulling back on exactly the homes buyers want
The single-family weakness is not random. It is a rational builder response to the same conditions squeezing buyers, and the mechanism is worth understanding.
Builders are contending with elevated mortgage rates and higher construction financing costs that continue to weaken confidence and demand, in the words of the homebuilders' association chairman. High rates hit single-family building twice: they raise the builder's own cost of financing construction, and they price out the buyers who would purchase the finished house, exactly the quarter-point qualification cliff squeezing buyers on the demand side. Faced with weak buyer traffic and swollen inventories of unsold new homes, builders have been cutting prices, subsidizing mortgage rates, and slowing construction of spec homes built without a buyer already lined up.
Multifamily runs on different fuel. Apartment demand is driven by renters, and renters multiply precisely when high rates and prices lock people out of buying. So the same conditions that suppress single-family construction can support multifamily, because the people who cannot afford to buy a house still need somewhere to live, and increasingly that somewhere is a rental.
What the report actually says
Put the pieces together and the June construction data is not the good news the headline implies. It is a picture of a housing market building fewer of the homes people buy and more of the ones people rent, with the forward indicators pointing down.
The genuine signal, underneath the 19% noise, is a third straight monthly decline in single-family starts, single-family permits near cycle lows, and builders telling you through their permit filings that they expect to build less. None of that resolves the core problem, that the country is not building enough of the moderately priced single-family homes that would actually improve affordability, and the report suggests it is about to build even fewer.
So the next time a housing-starts number "surges," the move is to ask one question the headline never answers: was it single-family or multifamily, and which way did permits go? In June, the answer to both turned a boom into a warning.
The number went up. The thing it was supposed to measure went down.