C — Q3 2026 Financial Report Analysis

Citi's Q3: EPS up 53%, but about half of that is last year's write-down dropping out

Citigroup earned $5.25 billion, or $2.85 per diluted share, in the three months to 30 September 2026, against $3.75 billion and $1.86 a year earlier. Revenue (net of interest expense) rose 8% to $23.85 billion and operating expenses fell 2% to $13.95 billion. The headline growth is inflated by the comparison: the year-ago quarter carried a $726 million goodwill write-down ($714 million after tax) tied to the planned sale of a stake in Mexico's Banamex. Without it, Citi's own figures put last year's profit at $4.5 billion and EPS at $2.24, so the like-for-like gains are +17% in net income and +27% in EPS. Those are still solid numbers. Against the record second quarter, though, the business slowed: revenue fell 4%, net income 10%, and return on tangible common equity dropped to 11.5% from 13.0%.

This analysis is based on Citi's earnings release and Quarterly Financial Data Supplement (Exhibits 99.1 and 99.2 to its 8-K of 13 October 2026). The full 10-Q for the quarter isn't filed yet, so management's discussion, the detailed credit tables and the notes are limited to what those two documents contain. Ratios as of 30 September are marked preliminary by Citi.

The quarter in numbers

Some of these terms need a short explanation. Net interest income is what Citi earns on loans and securities minus what it pays depositors and lenders. The net interest margin expresses that as a yield on interest-earning assets. The efficiency ratio is operating costs per dollar of revenue, so lower is better. Net credit losses are loans written off as unrecoverable, after recoveries. The CET1 ratio compares Citi's highest-quality capital with its risk-weighted assets. Regulators set the minimum, and that minimum also caps how much Citi can hand back to shareholders.

For the first nine months, revenue was $73.24 billion (+12%), net income $16.86 billion (+42%), diluted EPS $9.07 against $5.78, and RoTCE 12.6% against 8.6%.

Where the growth came from

Services produced about half of Citi's group income from continuing operations. Net interest income rose 17%, "driven by higher average deposit balances and deposit spreads." Fees rose 12%. Within that, cross-border transaction value was up 10%, commercial card spending up 10% and U.S. dollar clearing volume up 5%. Custody fees rose with a 16% increase in assets under custody and administration, which reflects both new clients and higher market values. Costs rose 10%, including severance and pay tied to performance, so the efficiency ratio improved by only 2 points.

Markets (trading) grew 6%. Equities rose 14% on prime brokerage, where Citi lends to and finances hedge funds: "prime balances up over 40%." Fixed income rose 4%, but only because of spread products (credit and financing), which grew 13%. Rates and currencies were flat: higher foreign-exchange volumes were offset by lower rates revenue. Costs rose 7%, so profit grew just 2%. Against Q2, Markets revenue fell 13% and its RoTCE fell to 12.3% from 17.0%. That is the main reason group returns dropped.

Banking grew 6%. Equity underwriting (ECM) jumped 77% "as the business gained share in follow-ons and convertibles." Advisory fell 4% "amid a lower wallet," meaning less merger activity across the industry. Debt underwriting fell 3%. Plain corporate lending revenue fell 16% on "lower loan spreads." Profit rose 42% largely because the provision dropped to $23 million from $157 million. Last year Citi built reserves as its loan book grew.

Wealth grew revenue 15%, and profit nearly doubled. Most of the gain came from deposits: net interest income rose 20% on "higher deposit spreads and average deposit balances." Investment fees rose 5%, and client investment assets grew 15%. Expenses rose only 3%, so the efficiency ratio fell to 75% from 84%. That is still the highest cost ratio of any Citi business, but it is falling fast.

U.S. Consumer Cards revenue was flat, and profit fell 8%. Net interest income rose 6% on larger balances, but fee revenue fell further below zero. Card revenue is reported after what Citi pays co-brand partners and spends on rewards, and that line grew 22% to $3.69 billion, while interchange fees (what merchants pay on each card purchase) grew 12% to $2.74 billion. That reflects the extra American Airlines co-branded portfolio Citi took on in Q2. The cost of the partnership shows up immediately, while the interest income on the balances builds over time. Spending on general-purpose cards rose 13% to $162.9 billion, and those loans grew 9%.

What the headline numbers hide

Most of the EPS jump comes from the comparison and from buybacks. Here is how 53% EPS growth breaks down:

• The $714 million after-tax write-down in Q3 2025 accounts for about half. Citi's own ex-item EPS growth is 27%.
• On that like-for-like basis, profit available to common shareholders rose about 17%, from roughly $4.19 billion to $4.89 billion (our calculation from Citi's figures).
• The diluted share count fell 8.4%, from 1,863 million to 1,707 million. That turned 17% profit growth into 27% EPS growth, so buybacks supplied about 10 of the 27 points.
• The tax rate didn't help. The effective rate fell to about 26% from 29%, but only because the write-down was barely tax-deductible. Excluding it, last year's rate was also about 26% (our calculation).

Two other drags grew between Citi's profit and its shareholders. The share of profit belonging to outside shareholders of subsidiaries rose to $175 million from $38 million, because others now own part of Banamex. Preferred dividends rose to $353 million from $274 million. Together that is $216 million more of each quarter's profit that no longer reaches common shareholders.

Interest income grew faster in trading than in lending and deposits. Markets net interest income rose 50% to $3.27 billion, while Markets non-interest revenue fell 20%. Citi says it manages trading on total revenue because positions are hedged across both lines, so that split is mostly accounting. The useful figure is net interest income outside Markets, about $13.71 billion, up roughly 7% (our calculation). That is the part earned on lending and deposits, and it grew about half as fast as the group total of 14%. Quarter on quarter, the group margin fell 11 basis points to 2.43%, and net interest income slipped 1%.

Credit got better, but reserves went up. Net credit losses were flat at $2.22 billion while loans grew 12%, so the loss rate fell to 1.09% from 1.21%. On general-purpose cards the loss rate fell to 3.51% from 3.85%, and loans 30–89 days late fell to 1.16% from 1.21%. Citi still built $369 million of reserves, against $236 million a year ago. In cards, it put the build "primarily" down to "changes in the macroeconomic outlook," not to its own borrowers getting worse. Even so, the allowance fell to 2.48% of loans from 2.65%, because the loan book grew faster than reserves did. A sizeable share of that loan growth is Markets lending to trading clients, not consumer lending.

One problem area: Mexico. Net credit losses in All Other rose 46%, "driven by higher volume and portfolio seasoning in Mexico Consumer." Seasoning means newer loans reaching the age at which defaults usually show up. The loss on the All Other line widened to $955 million from $756 million.

FX and costs. Currency moves added about $103 million to revenue growth, so growth was 7% excluding currency effects, against 8% reported. Excluding the write-down, expenses rose about 3%, driven by transaction costs, deposit insurance, pay and a contribution to the Citi Foundation. Lower professional-services costs, and lower "transformation" (restructuring) spending in All Other, partly offset those increases. Revenue still grew about 5 points faster than costs on a like-for-like basis (our calculation).

Capital is still going out faster than it comes in. CET1 held at 12.8% from Q2: net income and a "lower deduction from DTA and other" (deferred tax assets) offset buybacks, dividends and unrealized losses on securities. Common equity fell 1% year on year to $191.4 billion, and tangible book value per share rose 6% mainly because the share count fell 7%. The supplementary leverage ratio, a capital measure that ignores risk weights, fell to 5.0% from 5.2% in Q2 as the balance sheet grew to $2.94 trillion. Cash-flow conversion, a standard check for industrial companies, tells you little about a bank, where operating cash flow mostly reflects trading and funding flows. We leave it out.

Did last time's read hold up?

Our Q2 2026 analysis said Citi's 13.0% RoTCE needed to survive "a quarter when markets are quiet and reserves have to be built rather than run down." Q3 was that kind of quarter, and the answer is mostly no. RoTCE fell to 11.5%, Markets revenue fell 13% from Q2, and the reserve build rose to $369 million from about $100 million in Q2.

• Issuance: we expected a slower issuance quarter to show up in Banking's 63% efficiency ratio. It did: Banking revenue fell 9% from Q2 and the ratio rose to 70%. Equity underwriting was still up 77% year on year.
• Cards: we flagged rising 30–89-day delinquencies as the new American Airlines accounts age. That hasn't happened yet: early delinquencies on general-purpose cards edged down to 1.16% from 1.17%, and loss rates fell. The partnership's cost did show up: partner payments and rewards rose another $129 million from Q2.
• Banamex and buybacks: both went as committed. The final 1.4% Banamex stake sale closed in Q3, and capital return continued at $4.1 billion.

Takeaway: Strip out last year's Banamex write-down and Citi grew EPS 27% on 17% profit growth, with buybacks supplying the difference. The repeatable engine is clear: Services earns a 31% return on deposit-driven growth, and Wealth is fixing its costs. But the drop from 13.0% to 11.5% RoTCE in one quieter quarter shows that the 11–13% goal still depends on trading and capital markets. Interest income from lending and deposits is growing about 7%, not the 14% the group total suggests.

What to watch

Citi gave no new numerical guidance in the release. Jane Fraser said Citi expects "to exceed this year's return target, even as we pull forward investments toward our 11-13% near-term goal," with a longer-run aim of 14–15%. The release doesn't restate this year's target figure. Year-to-date RoTCE is 12.6%. Citi lists its risks as "higher interest rates, elevated inflation and conflicts in the Middle East."

Our view on trajectory:

• Q4 is seasonally Citi's weakest quarter. In Q4 2025, RoTCE was 5.1% and revenue $19.87 billion, including a large negative "other revenue" line. Beating the full-year target from 12.6% year to date looks likely. A double-digit Q4 return would be the real test of the 11–13% goal.
• "Pulling forward investments" means costs will rise. Expenses fell 2% from Q2 this quarter. Management is now telling investors that spending will accelerate, so the efficiency gains of 2026 may narrow in 2027.
• Banamex's currency-translation charge still lies ahead. The Q2 10-Q said Citi may book about $9 billion of currency-translation losses (losses from the peso's past moves, held in equity until now) and deconsolidate Banamex in early 2027. The charge is non-cash and, per Citi, capital neutral, but it will make one quarter's reported earnings look very poor. The Q3 10-Q should update the timing.
• Cards credit as the AA book ages. Loss and delinquency rates improved this quarter. Whether they hold as the acquired accounts season will decide whether the Cards segment's 21.8% return survives a rising partner-payment bill.

Next report: Citigroup's full 10-Q for Q3 2026, with segment MD&A, credit tables and the Banamex update, is expected in early November 2026 (around 4 November, our estimate from past filing dates). We will update this analysis if it changes any figure here.