A 30% jump in mortgage volume sounds like unambiguous good news. Read the fine print of the second-quarter bank earnings and it splits into two questions that the headline number cannot answer: is anyone actually making money on that volume, and where did it come from.
The raw numbers were genuinely strong. JPMorgan Chase reported $17.2 billion in origination volume, up 26% from the first quarter and 27% from a year earlier. Wells Fargo's $9 billion was up 43% sequentially and 22% year over year. Bank of America's consumer division produced $3.5 billion in first mortgages plus $2.4 billion in home equity. Combined, the three posted a gain in the 30% range, according to notes from BTIG and Keefe, Bruyette & Woods.
That crushed the forecasts. The Mortgage Bankers Association's June outlook had called for a 3% industrywide quarter-to-quarter gain; even Fannie Mae's more bullish number was around 9%. The banks came in at roughly ten times the MBA's estimate. When actual results are an order of magnitude above the consensus forecast, that is not a beat. That is the forecast, and possibly the story, being wrong about something.
The number that complicates the celebration
Here is the figure the upbeat framing has to walk past. Only Chase still discloses its gain-on-sale margin, the spread that determines how much profit a lender actually books on each loan it sells. Chase's came in at 85 basis points, about 45 basis points lower than three months earlier.
Sit with the shape of that. Volume up roughly 26% at Chase, margin down roughly 35% quarter over quarter. Those two facts point in opposite directions, and the second one is the one that pays the bills. A lender can originate a record pile of loans and make less money doing it if the margin on each loan collapses. Volume is the number in the headline. Margin is the number in the bank account.
KBW's Bose George flagged exactly this. He judged Chase's margin worse than expected, pointedly noting it happened during a seasonally strong quarter for volume. In other words, this was the good part of the year, the stretch when margins should hold, and the one visible margin still cratered. That is not what a healthy, broadly profitable boom looks like.
The honest caveats belong here too. Bank gain-on-sale margins are historically far more volatile quarter to quarter than nonbank margins, so one bad print is not a trend. And George noted the drop could reflect pipeline hedge losses or other one-time items rather than genuine pricing pressure. Fair. But that is precisely the problem with the whole quarter: we cannot tell, because the disclosure that would let us tell has mostly disappeared.
The disappearing scoreboard
Buried in the reporting is a structural fact that should bother anyone who relies on these earnings to understand the mortgage market. Chase is now the only one of these three banks that discloses gain-on-sale margin at all. As George put it, it is hard to spot trends when a single bank is the last one publishing the number.
This is a slow erosion of transparency that nobody voted on. Bank of America does not break out mortgage banking income at all. The banks disclose the flattering figure, volume, loudly, and have quietly stopped disclosing the figure that reveals whether the volume is profitable. The result is an information environment where the industry can look booming on the one metric everyone publishes while the metric that would show strain goes dark. When only one company still shows its work, the market's read on profitability rests on a sample size of one, and that one just posted a sharp decline.
The question that reframes everything
The sharpest line in the analyst commentary is the one that quietly dismantles the celebration. KBW's George wrote that the quarter looked positive for volume, but that it remains unclear whether it reflects industry volume or banks taking share from nonbanks.
That distinction is the entire story, and it changes the meaning of the 30% completely.
If the banks' surge reflects a rising overall market, it is good news for everyone, banks and nonbank lenders alike, and the boom is real. If instead the banks are pulling business away from nonbanks in a market that is barely growing, then the same 30% is not expansion. It is a transfer, and it is actively bad for the independent mortgage banks who dominate US originations. One reading is a rising tide. The other is three of the largest, best-capitalized players in the country using scale and pricing to take share, which would explain the collapsing margin: you buy market share by cutting price, and cutting price is what a 45-basis-point margin drop looks like.
The combined bank volume gain also outran the quarter's 11% rise in mortgage-backed securities issuance. If the total pool of loans being securitized grew far less than the banks' originations did, that is circumstantial support for the share-grab reading. The banks grew much faster than the market they operate in, which usually means the growth came from somebody else's book.
Why the analysts still cheered
Given all that, it is worth asking why BTIG's Doug Harter called the results a positive sign for the more production-focused nonbanks in his coverage, naming Rithm Capital and United Wholesale Mortgage as top picks.
The bullish logic runs like this: if big, slow-moving banks with weaker mortgage technology can post 30% volume gains, then the nimble, production-focused nonbanks should do at least as well, because demand strong enough to lift the banks should lift everyone. It is a reasonable inference. But it quietly assumes the answer to George's question, that the volume reflects a growing market rather than share moving between players. If the banks grew by taking from nonbanks, then strong bank volume is a warning for nonbanks, not a tailwind. The same data point supports opposite conclusions depending on an unknown nobody has resolved. Harter himself only projected a 3% volume increase for his nonbank coverage, a striking gap from the banks' 30%, which is itself a small piece of evidence that the two groups may not be riding the same wave.
What to actually watch
Set aside the headline. The 30% is real and it is not the point. Three things will tell you what the quarter meant, and none of them is the volume figure.
First, the nonbank results still to come. If Rocket, UWM, Rithm, and PennyMac post volume gains anywhere near the banks', the rising-market story holds. If they come in near Harter's 3%, the banks took share, and the boom was partly a reshuffling. Second, margins across the board, to the limited extent they are still disclosed, because volume without margin is motion without profit. Third, whether more disclosure disappears, because a market where the last bank reporting gain-on-sale stops doing so is a market getting harder to read by design.
The banks had a big quarter. Whether the mortgage industry did is a genuinely open question, and the one visible margin number suggests the answer is more complicated than 30% makes it sound.