Companies & Markets October 8, 2026

The October Print

For the big lenders, October has been the strongest earnings window of the year: 14 of 16 since 2010. This year the 10-year sits at 5.3%, a level with no precedent anywhere in the data.

Alphatica Research · BAC, JPM, WFC, C, GS, MS

Point-in-time note. All prices, estimates and consensus figures are as of October 8, 2026 and are not updated after publication.

Since 2010, BAC, JPM and WFC beat the S&P 500 from the October print to the January print in 14 of 16 years, by about 7% to 8.5% on average. We are halving that this year.

Base rate

+3% to +5%

Versus the S&P 500, October print to January print, for BAC, JPM and WFC. Four separate approaches land in that range.

Why halved

10-year at 5.31%

Our history covers October prints with the 10-year between 0.8% and 4.7%. Today's level is outside the sample entirely.

16 October prints, 2010 to 2025 · measured first close after earnings to the day before the next report

Summary

The short version

The selloff is not the signal. All six big banks beat Q2 estimates and fell about 9% afterward. How banks trade into the print has told us nothing about what happens next.
The print-day move is noise. The edge shows up in the weeks after earnings, not on the day.
It is not a rates trade or a sector trade. Rates explain only 11% to 15% of the pattern, and it holds against other financials.
What the print says matters, but only in October. Bigger beats have meant bigger moves into January.
The fundamentals are strong but stretched. Deal pipelines and credit look good. Margins and capital are under pressure from higher rates.
GS and MS have no reliable October edge.

The Setup

Every bank beat, the market sold them anyway

All six banks beat Q2 estimates in July. All six are lower three months later. The biggest beats had the worst follow-through.

Bank	Q2 beat	Day-1 move	Since release
GS	+45.0%	+9.00%	-14.22%
MS	+19.7%	+0.39%	-16.10%
C	+15.8%	-5.29%	-8.66%
WFC	+13.3%	-2.71%	-7.03%
JPM	+9.8%	+2.50%	-0.97%
BAC	+7.1%	+1.88%	-9.09%
Average		+0.96%	-9.34%

Price return from the last close before each report to the October 6 close.

The beats were real. US banks posted their best profitability in a decade, and JPM's trading revenue more than doubled to $12 billion. But part of Q2 was one-off: the $75 billion SpaceX IPO paid each of these banks $75 to $100 million in fees, and management at JPM and MS said Q2 included events unlikely to repeat.

Meanwhile the 10-year climbed from about 4.6% to above 5.2%. The question for October: is the discount an opportunity or a warning?

The History

What 16 years of October prints say

We measured every quarterly print from 2010 to 2025, from the first close after earnings to the day before the next report. October ranked first of the four quarters for all six banks.

Bank	October drift vs. S&P 500	Positive Octobers	Significant?
WFC	+8.49%	81%	Yes
BAC	+8.43%	88%	Yes
JPM	+7.31%	88%	Yes
MS	+6.82%	81%	Yes
C	+6.40%	75%	Borderline
GS	+6.11%	75%	Borderline

On the print day itself there is nothing. JPM rose on only 4 of its 16 October print days.

Does it hold up?

Three years had huge November rallies inside the window: the 2016 election, the 2020 vaccine news and the 2024 election. We removed them, then also measured against other financials (XLF) instead of the S&P 500.

Bank	All years	Without 2016, 2020, 2024	vs. financials
WFC	+8.5%	+4.8%	+4.9%
BAC	+8.4%	+4.4%	+5.4%
JPM	+7.3%	+4.0%	+3.8%
MS	+6.8%	+2.8%	+4.0%
C	+6.4%	+1.5%	+3.1%
GS	+6.1%	+1.8%	+3.3%

Gold = still statistically significant. Only the three big lenders survive both tests. That is the core of the pattern.

What drives it

We tested each obvious explanation:

Interest rates? Mostly no. Yield moves explain only 11% to 15% of the effect, and the level of rates going in did not predict the outcome.
Financials doing well in Q4? No. The lenders still beat the sector.
Just the time of year? No. The same window shifted three months earlier or later shows nothing, or turns negative.
What the print says? Partly. In October only, bigger beats led to bigger runs into January. In other quarters, beat size did not matter.
The day-one reaction? No. The market's first move does not predict what follows.
A selloff into the print? No. Neither bounce-back nor continued weakness shows up in the data.

Lender October drift vs. S&P 500	Average	Positive
Big beat	+12.8%	75%
Middle	+5.2%	75%
Small beat or miss	+6.2%	81%

Even small beats and misses drifted higher on average. A big beat roughly doubled the result.

When rates were higher

October still worked when the 10-year was above 3%, at about half the strength: +4.8% on average across 2018 and 2022 to 2025. The closest setups to today were 2023 (+8.7%), 2024 (+12.5%), 2022 (-0.8%), 2025 (+3.0%) and 2018 (+0.5%). None had a 10-year above 4.7%.

The Anomaly

Why 2026 is different: the 10-year at 5.3%

Our history covers October prints with the 10-year between 0.8% and 4.7%. Today it is 5.31%, up 53 basis points in a month. It has closed above 4% on 188 of 2026's first 189 trading days. The Fed hiked in September, and markets are pricing better-than-even odds of another hike on October 28.

October	10-year going in	1-month change
2026	5.31%	+53 bp
2025	4.05%	+1 bp
2024	4.09%	+44 bp
2023	4.70%	+43 bp
2022	3.97%	+55 bp
2018	3.14%	+16 bp
2012 to 2021	0.79% to 2.71%	-25 to +24 bp

In the past, higher yields helped the lenders. Today they are hurting them: since July the 10-year rose about 70 basis points while the lenders lagged the S&P 500 by about 10%. In 47 of 48 past October prints, the lenders were trading with yields going in. Today's setup has almost no precedent.

Four ways high rates reach bank earnings

1. Deposit costs

The risk is that banks have to pay depositors more with each hike, squeezing margins. Margins have already slipped from late-2025 levels at five of six banks.

2. Interest income

The commercial banks still earn more from a further hike, but the gain is shrinking. A 1-point rise adds about $1.8 billion a year at JPM, $1.3 billion at WFC (down from $1.9 billion at year-end), $1.2 billion at C and $1.0 billion at BAC. Falling rates would hurt more than rising rates help.

3. Capital

Rising long-term yields cut the value of banks' bond holdings, which eats into capital and buyback room. Citi estimates a 1-point rise costs about $3 billion after tax. BAC, with the largest legacy bond book, carries the same risk.

4. Commercial real estate

About $875 billion of commercial mortgages mature this year, roughly half on bank balance sheets, and new loans cost about 1.5 points more. Charge-offs are still falling, so this risk is ahead, not yet in the numbers.

Bank	Margin, Q2 2025	Margin, Q2 2026	Change
WFC	2.65%	2.38%	-27 bp
JPM	2.20%	2.16%	-4 bp
C	2.34%	2.31%	-3 bp
MS	1.08%	1.06%	-2 bp
BAC	1.83%	1.94%	+11 bp
GS	0.73%	0.77%	+4 bp

Net interest margin.

Bank	Net charge-offs, Q2 2026	Q2 2025
C	1.17%	1.22%
JPM	0.59%	0.68%
BAC	0.44%	0.52%
WFC	0.34%	0.43%
GS	0.30%	0.44%
MS	0.03%	0.02%

Who is most exposed

WFC: the sharpest margin squeeze, though management says Q3 is tracking better than guided.
JPM: the most exposed if deposit costs suddenly jump. Its CFO flagged that risk directly.
BAC: capital pressure from its bond book is the key watch.
C: capital pressure plus the highest card losses in the group.
GS and MS: less exposed to margins, more exposed to any slowdown in deals.

Here the history and the fundamentals disagree. The data says rates going in did not predict October outcomes. But sudden deposit repricing and capital stress are risks that may only appear above the levels history covers. That is why we halve the base rate.

The Bull Case

The case for another beat

Management teams were upbeat at September's industry conferences.

Deals: JPM guided Q3 investment-banking fees and trading revenue up mid- to high-teens. GS entered Q3 with its biggest deal backlog in five years. MS has a record 100-plus private-equity sale mandates and calls the deal cycle early to mid innings.
Consumer credit: card losses fell to their lowest since July 2023. BAC's CEO called credit quality as good as it has been in a long time.
Interest income: JPM raised its full-year guide to $105.5 billion. BAC expects the top of its 6% to 8% growth range. WFC says Q3 margins are tracking better than guided.
Capital return: all six passed the 2026 stress test with the lowest projected losses in at least seven years. GS, MS, JPM and C raised dividends. C launched a $30 billion buyback.

Bank	Conference tone
JPM	Bullish
MS	Bullish
C	Positive
WFC	Constructive
BAC	Mixed-positive
GS	Cautiously positive

The offsets: Q3 will be below Q2's exceptional quarter. GS flagged more than $500 million of higher expenses. JPM and WFC expect deposit costs to creep up. And on September 22, fears that AI agents could pull deposits away from banks knocked 2% to 4% off the group in a day.

Expectations

What the Street expects

Analysts are most bullish on the same three lenders our history favors, and least bullish on GS.

Bank	Consensus rating	Price target	Upside
BAC	Outperform (1.56)	$67.33	24.5%
C	Outperform (1.65)	$152.75	18.9%
WFC	Outperform (1.78)	$99.20	21.7%
JPM	Outperform (2.12)	$372.81	12.5%
MS	Outperform (2.28)	$223.27	16.9%
GS	Hold (2.60)	$1,079.50	20.3%

1 = Strong Buy, 3 = Hold, 5 = Strong Sell. As of October 7.

Bank	Q3 2025 EPS	Q3 2026 consensus
JPM	$5.07	$5.89 to $6.06
GS	$12.25	$13.91 to $14.43
WFC	$1.73	$1.83 to $1.92
BAC	$1.04	$1.10 to $1.16
C	$2.26	$2.85
MS	$2.80	$3.05

Ranges reflect different estimate providers. C's Q3 2025 actual excludes a one-time loss on its Mexico unit.

This matters because the size of the beat is the one print-day number that has predicted October follow-through. Q2 beats of 7% to 45% suggest estimates have been set conservatively.

Names

Bank by bank

Bank	October edge	Since July vs. S&P 500	Street view
BAC	+4.4%	-14.2%	Most bullish
WFC	+4.8%	-7.8%	Bullish
JPM	+4.0%	-6.8%	Bullish, more Holds
C	+1.5%	-7.0%	Bullish
MS	+2.8%	-19.4%	Divided
GS	+1.8%	-24.8%	Only Hold

Average drift vs. the S&P 500, excluding 2016, 2020 and 2024. Gold = statistically significant.

BAC. The strongest October record, positive 88% of the time, and the most bullish analysts. Also the deepest lender selloff. Watch capital pressure from its bond book.

WFC. The largest October gains on record and the largest remaining edge. The sharpest margin squeeze, but Q3 margins are tracking better than guided.

JPM. The most statistically reliable October pattern, but the weakest print-day record. Strongest guidance in the group, and the most exposed if deposit costs jump.

C. Positive history that is not reliable once the big years are removed. Highest card losses, offset by a $30 billion buyback.

MS. Marginal history and a deep selloff. Well positioned for higher rates, but management warned about market froth.

GS. No reliable October edge, the biggest selloff and the only Hold rating. The record backlog is real, but the stock has given back its entire Q2 beat and more.

Context

The global picture

The beats were not just a US story: 21 of Europe's 25 largest banks beat Q2 estimates. The biggest came from banks with trading desks or recent acquisitions, including ABN AMRO (+16.6%), SEB (+13.2%), NatWest (+13.0%), Société Générale (+12.2%) and UBS (+10.7%). Trading volatility and one-off deals drove much of it, and those are the drivers least likely to repeat in Q3.

Watch List

What to watch

Our base case for BAC, JPM and WFC: +3% to +5% versus the S&P 500 from the October print to the January print. Four different approaches land in that range.

Approach	Estimate
Octobers with the 10-year above 3%	+4.3% to +5.2%
Closest past setups	+4.8%
Low-rate years without the big rallies	+5.1%
Forward-looking model	+2.5% to +3.8%

When the numbers drop:

The size of the EPS beat. It is the one number that has predicted October follow-through. A big beat has roughly doubled the move.
2027 commentary. October is usually the first look at next year's interest income, expenses and buybacks.
Deposit costs and margins. A jump at JPM or WFC would be the clearest sign high rates are breaking the pattern.
Credit. Any turn in commercial real estate or card losses.
Capital. Bond-book losses at BAC and C, and balance-sheet limits at GS.

What not to over-read: the day-one move, and the selloff going in. Neither has told us anything about January.

The October print has been the most reliable seasonal window the big lenders have. Fourteen of sixteen years, roughly 8% against the index, and it survives the removal of the three election and vaccine rallies that would otherwise flatter it.

It has never been tested with the 10-year at 5.3%. That is the whole argument for halving the base rate rather than abandoning it: the pattern is real, the sample does not cover today, and the honest answer to an out-of-sample setup is a smaller number, not a louder one.

Methodology. Drift is measured from the first close after each quarterly report to the close before the next report, relative to the S&P 500, across 16 October prints from 2010 to 2025. Statistical significance is assessed on the distribution of those 16 observations and reported both for the full sample and excluding 2016, 2020 and 2024, the three years containing outsized November rallies inside the window. The sector robustness check substitutes XLF for the S&P 500 as the benchmark. Rate attribution regresses drift on contemporaneous 10-year yield changes. The forward-looking estimate blends the rate-conditioned subsample with current positioning and consensus revisions. Historical base rates describe what has happened in a limited sample. They are not forecasts, and the current 10-year level sits outside the range the sample covers.