Banorte earned a 25.7% return on equity in the second quarter, and its banking arm cleared 32.4%. Those are the numbers a bank posts once a cycle, if it is lucky. On Friday, the day Mexican miners ran roughly 5% to 6% and pulled the IPC to its best week in months, Banorte gained 1.0%. The most profitable bank in the country, coming off the most profitable quarter it has reported in years, could not get investors to look up from copper.
That gap, between what the banks earned and what the market will pay for it, is the story. Mexico’s listed lenders just closed a Q2 earnings season that looks, on the surface, close to perfect: record margins, mid-20s returns on equity, loan books still growing. And the sector trades at roughly eight times earnings, has sat out this month’s rally, and cannot get a bid. The market is not confused. It is pricing these earnings as a peak rather than a run rate, and the evidence inside the sector says it is right.
The margin is Banxico’s, not Banorte’s
Start with where the profit comes from, because it is not the franchise. Banorte’s bank-level net interest margin hit 6.9% in the second quarter, above the top of its own guidance. Group net income reached MXN 15.6 billion (~USD 923mm at 16.90 MXN/USD), lifting first-half earnings to MXN 31.0 billion, up 4% year over year. The return on equity, 25.7% at the group and 32.4% at the bank, sits near the ceiling of anything Banorte has printed.
None of that is a mystery, and none of it is skill. It is the 6.50% policy rate. We laid out the mechanism two weeks ago in “Inflation Hit Target. Banxico Won’t Move. The Peso Is the Reason.”: a central bank holding a nominal 6.50% against 3.1% inflation runs a real rate near 3.4%, and every month it declines to cut, that spread stays wide. Banks fund themselves in cheap current-account deposits and lend at rates pinned to a policy rate that has not moved since May. The difference drops to the bottom line. Banorte’s own full-year guide tells on itself: management sees the bank margin holding around 6.8% to 6.9%, flat from here, and net interest income up 15% for the year. That is a margin at its ceiling and an income line still riding the carry. To keep that income growing once the spread stops widening, the bank is leaning into its highest-yielding, highest-risk lines. Its 2026 guidance calls for consumer credit up 10% to 14%, credit cards up 14% to 18%, and auto loans up 15% to 20%, all running well ahead of the 6% to 7% pace it guides for mortgages. That is a deliberate reach for yield at the top of the rate cycle, and it sets up the cost-of-risk question we come back to below.
The tell is what happens to this engine when the rate finally falls. Banxico’s minutes, released August 20, describe a long pause with the output gap still negative and convergence to the 3% target pushed out to late 2027. The hold is the banks’ friend right up until it ends, and the day it ends, the same arithmetic runs in reverse. A record margin built on an unchanged policy rate is not a moat; it is borrowed from the central bank on terms the market can read.
Records at eight times earnings
Here is the disconnect in one line: the most profitable quarter Mexican banks have had in years bought them nothing on the tape. Banorte trades near eight times forward earnings. On the week the IPC clawed back to 65,224, up 1.4% on Friday and out of correction, the move was Grupo México up 6.2%, Peñoles up 5.2%, Cemex up 3.9%. Banorte managed 1.0% on a fraction of the turnover, and it had slipped 1.5% on the previous mining-led session. Walmex dropped 2.1%. The bid went to copper and silver, to companies selling into record metal prices, and not to the bank compounding capital at a 25% return.
Read what that pricing says. The market is happy to chase miners at the top of a commodity cycle because it treats the metal price as the risk it is paid to take. It will not extrapolate a 25% bank ROE because it treats that number as borrowed from a rate setting that reverses. A cyclical peak in copper gets a premium. A cyclical peak in bank margins gets a discount. The two calls are consistent only if you believe, as the tape clearly does, that Mexican bank profitability is closer to its high than its base.
That is a defensible read, not a cheap one. Bank multiples in Mexico have been compressed for years on political risk, judicial reform, and the nearshoring story that keeps arriving as capital and not as jobs, a thread we followed in “The Capital Showed Up. The Jobs Did Not.” An eight-times multiple on a 25% ROE is the market telling you it does not trust the E to hold. The earnings kept climbing through the first half and the multiple did not follow, which is a market refusing to pay up for a number it expects to fade. When the cheapest, most profitable part of the index is the part nobody will buy, the market has already priced a view on where the earnings go next.
Banregio already blinked
If Q2 were the start of a durable earnings run, you would not expect the bank closest to the real economy to be cutting its numbers. Banregio did exactly that. Alongside its second-quarter results, the Monterrey lender, listed on the BMV as Regional under the ticker RA, took down full-year guidance across the board: net interest margin to a 5.6% to 6.0% range from 6.0% to 6.3%, return on equity to 16.5% to 17.5% from 18% to 19%, and net income growth to somewhere between minus 5% and flat, from a prior 5% to 10%. Margins compressed in the quarter even as the loan book grew 11%. That is a bank running faster to stay in place.
Banregio matters here because of what it is: a mid-size, business-focused lender without Banorte’s scale, diversification, or fee engine to paper over a thinner spread. It feels the compression first and cannot hide it. And it is not alone at the smaller end. Banco del Bajío, the Bajío region’s SME and agricultural lender, missed on earnings, printing MXN 1.82 per share against a MXN 1.90 estimate. Inbursa saw second-quarter revenue fall 14% year over year. The further you get from Banorte’s diversified balance sheet, the more the Q2 numbers stop looking like a boom and start looking like a plateau with cracks in it.
This is the same two-faced setup we described for the whole economy on August 10, now visible inside one sector. Banorte’s record is the trailing indicator, the carry already banked. Banregio’s guidance cut is the leading one, the first bank to say out loud that the spread has stopped widening. The sector is not homogeneous, and the tiers are telling different stories at the same time.
6.9% NIM
+13% book
was 18 to 19%
vs 1.90 est.
-14% y/y
What the fat margin is lent against
A wide margin flatters a bank only if the loans behind it get paid. That is the question the headline ROE does not answer, and it is where the K-shaped economy comes in. System credit growth has already slowed to roughly 3.8% in real terms as of the first quarter, and consumer lending decelerated from 13.5% real growth at the end of 2024 to 8.8% a year later. Banks are earning more on a book that is growing less. S&P sees total commercial-bank credit expanding only 6% to 8% this year, so the volume tailwind is modest before the consumer book even turns. Consumer delinquency has started to tick up, and the CNBV has flagged it, tracing the rise to an economy running below potential, the drag of past inflation, and thinner public transfers. The total non-performing ratio, 2.4%, is still low. The direction is the concern, not the level.
Now put Gentera in that picture. The microlender posted a 41.2% net interest margin in the second quarter, a 24.5% ROE, and a loan book up 13% to MXN 94.6 billion (~USD 5.6bn), serving 6.7 million mostly informal borrowers across Mexico and Peru. First-half net income of MXN 4.8 billion was a record. We argued in “Gentera FY 2025: Loan Book Growth Is Slowing. Is That Actually Bad?” that the company’s earlier deceleration was healthy discipline, not weakness, and the reacceleration since is real. But a 41% margin is not free money. It is the price of lending to the exact borrowers, the bottom of the K, that a manufacturing sector shedding workers for over a year squeezes first. That margin is risk compensation for a cost of risk that has not fully arrived. As long as those loans perform, Gentera prints. The moment the informal economy softens, the same book that produces a 41% margin produces the losses that margin was pricing.
None of the listed banks are in trouble. Banregio’s cost of risk held near 0.4% in the quarter with non-performing loans around 1.4%, and capital ratios across the sector are strong. The point is subtler and it cuts against the headline: the profitability is highest exactly where the economy is weakest, and the fat margin is a warning about the borrower as much as a reward to the shareholder.
The asymmetry
Weigh both sides honestly, because the banks are not a short and the multiple is genuinely cheap.
Bull case. Eight times earnings on a 25% ROE is a low bar to clear. Banorte is returning capital, paid a USD 0.60 per share dividend in May, and generates more than a slow-growing loan market can absorb, so the excess comes back to shareholders. If Mexico lands softly, the rate stays higher for longer as Banxico’s own minutes suggest, and loan growth stabilizes, these banks compound book value at mid-20s returns and pay you to wait. Unloved and cheap is where re-ratings start.
Bear case. The earnings trajectory bends down from here, not up. The first Banxico cut compresses the margin that produced the whole record. Loan growth is already slowing, so the volume cannot offset the spread. And the cost of risk clock is running in the consumer and micro books, where the K-shaped bottom lives. Pay a peak-cycle multiple on peak-cycle earnings and you can be right on the business and still lose on the stock.
Key risks. The banks are already an out-of-favor trade, so the pain may be less about a crash than about dead money: a sector that earns beautifully and goes nowhere because every strong print is discounted as the top. The swing factors are the timing of the first cut, the path of consumer delinquency, and whether nearshoring finally converts into the payrolls that would give loan growth a second leg.
The bottom line
Mexican banks just reported a quarter that, in isolation, reads like the best of a cycle: Banorte at a 25.7% ROE and a 6.9% margin, Gentera at a record, returns most sectors would envy. The market’s response was to buy copper. That gap is not a mispricing waiting to be corrected; it is the market pricing bank earnings as a peak, with the sector handing over the evidence itself: Banregio cutting guidance, Bajío missing, Inbursa’s revenue falling, credit growth slowing, consumer delinquency turning up. The record is the carry already collected. The guidance cut is the bill starting to arrive.
The through-line to everything we have written since June holds. The 6.50% hold that keeps the peso bid near 16.90, the strongest in over a year, is the same hold that fattens these margins, and it is the same hold that throttles the economy the loans are written against. Q2 was that trade at its most flattering, the moment the banks monetized the rate before anything came due. The reckoning arrives first at the smaller banks and in the consumer book, and Banregio has already given the tell. Until then, a record quarter at eight times earnings is not a bargain the market missed. It is a peak the market can see.