The KBW Bank Index Fell 9.7% in a Month. JPMorgan's Trading Desk Grew 86% in the Same Stretch.

Five of the largest banks in the United States report third-quarter results this week, and their stock prices have already told part of the story before a single number comes out. The KBW Nasdaq Bank Index, which tracks 24 lenders, dropped 9.7% over the past month while the S&P 500 gained roughly 1.5% over the same stretch. Bank of America fell 13.3%. Goldman Sachs fell 13.1%. Morgan Stanley dropped 11.7%. Charles Schwab and PNC each lost more than 10%. By late September the index had already fallen into correction territory, down more than 10% from its August 17 high.

None of that happened because trading desks went quiet. Three months ago, these same banks posted some of their strongest trading results in years. JPMorgan's markets revenue rose 35% to $12.1 billion last quarter, with equities trading up 86% to $6.0 billion. Goldman's equities business grew 72% to $7.42 billion, helping lift net revenue 39% to $20.34 billion. Morgan Stanley posted record equity trading revenue of $6.3 billion on the way to a 26.6% return on tangible equity. A trading boom this strong would normally lift bank stocks. This time it hasn't.

[[IMG1]]

What Changed

JPMorgan, Goldman Sachs, Wells Fargo and Citigroup report Tuesday morning. Morgan Stanley follows Wednesday. Wall Street's consensus for JPMorgan calls for $5.82 a share on $51.21 billion of revenue, up from $5.07 a share a year earlier, and the bank cleared a similar bar comfortably last quarter, posting $6.14 against a $5.79 estimate on revenue of $58.02 billion versus $50.19 billion expected. Citigroup's consensus sits at $2.62 a share.

The guidance heading into the quarter points toward another strong trading period. JPMorgan co-president Doug Petno told investors at a Barclays conference on September 15 that both markets revenue and investment banking fees should grow in the mid-to-high teens percentage range in the third quarter. Bank of America gave a different signal on one line of business, guiding its own investment banking fees down roughly 10%, a reminder that the deal environment hasn't moved in one direction for everyone.

What's actually changed since the second quarter is the backdrop these results now arrive against. The 10-year Treasury yield climbed as high as 5.31% this month, its highest level since 2000 at a recent auction, and is up roughly 1.15 percentage points since January. The Federal Reserve raised its target rate in mid-September, and futures markets still assign roughly 70% odds to another hike in December. The two-year to 10-year spread has narrowed to its tightest level since March 2025, the kind of flattening that squeezes the margin banks earn between what they charge on loans and what they pay on deposits.

Why It Matters

A flatter curve hits banks in a specific, mechanical way. Net interest margin narrows when short-term funding costs stay elevated while long-term loan yields don't rise fast enough to offset them. Treasury bills now compete directly with savings accounts for the same money, pushing up what banks have to pay to keep deposits from walking out the door. Slower loan demand compounds the problem: higher rates discourage the mortgage and consumer borrowing that gives banks steady income between trading cycles.

Dealmaking has cooled at the same time. North American M&A volume fell 23% in the third quarter, a headwind for the advisory fees that usually complement a good trading quarter. A newer worry has entered the conversation too. Apollo's chief economist, Torsten Slok, has warned that AI shopping agents, such as Meta's recently launched Muse, which can help a customer cancel a subscription or hunt down a better rate, could erode the inertia that lets banks keep paying below-market rates on checking and savings balances. Depositors who can be nudged toward a better rate as easily as asking an AI agent make for a less dependable, cheaper source of funding than the one banks have counted on for decades.

These figures don't read like a bad quarter on paper. Every one of the five banks beat estimates in the second quarter, and the guidance into the third points to more of the same. The disconnect is between results that keep beating expectations and a stock group that fell into correction territory anyway, which is usually a sign investors are pricing in something further out than next week's earnings print.

What to Watch Next

The reports land Tuesday and Wednesday morning, and the numbers that matter most might not be the headline EPS beats. JPMorgan built its 2026 outlook around roughly $105.5 billion in net interest income and an adjusted expense base near $107.5 billion; whether that guidance holds as the yield curve keeps shifting is worth tracking line by line. Deposit costs and loan growth commentary on the earnings calls deserve attention too, since executives will likely face questions about whether the AI-agent threat to deposit inertia shows up yet in actual numbers or only in warnings like Slok's. Trading revenue itself is the cleanest test of Petno's guidance: mid-to-high teens percentage growth is the bar, and the back half of the quarter, when yields accelerated toward 5.3%, is when that bar gets hardest to clear.

Credit quality is worth a glance as well. JPMorgan guided to a card net charge-off rate near 3.2% for 2026, and any meaningful uptick from that level would suggest consumers are feeling the higher-rate environment more than the trading headlines let on.

The Pulse24 Take

The gap between what these banks are earning and what their stocks are doing looks less like a verdict on the quarter about to be reported than a bet on what comes after it. Trading desks have had a historically good run, and nothing in the guidance so far suggests the third quarter breaks that streak. What the market seems to be pricing instead is the slower-moving stuff: a curve that's flattening again, deposit funding that may not stay as cheap as it's been, and loan demand that higher rates keep dampening. Those pressures rarely show up in a single earnings beat. They build over several quarters, which means investors who sold bank stocks into a 9.7% monthly decline are effectively betting that bill arrives before the trading boom fades. This week's five reports won't settle that bet in either direction, but the commentary on deposit costs and net interest margin will say more about 2027 than any single EPS number will.