Every credit union ALCO deck built since March opened with the same assumption: the Fed would be cutting by year-end — the only debate was how many. On Wednesday afternoon, Kevin Warsh's first Fed dot plot turned that assumption into wallpaper.

The Federal Open Market Committee held the federal funds rate steady at 3.50%–3.75% on June 17, 2026 — its fourth consecutive hold. The headline was a yawn. The Summary of Economic Projections was not. The median dot for year-end 2026 jumped to 3.8% from 3.4% in March — now implying at least one rate increase. The committee split almost evenly to get there: of the 18 participants who submitted projections, nine penciled in a hike, eight saw no change, and one a cut. As recently as March, the same committee was signaling the opposite.

Three months ago it was telegraphing cuts. Now it's telegraphing tightening. That is not a marginal revision. That is a pivot.

What actually changed

The inflation forecast did the damage. The June SEP lifted the median PCE inflation projection to 3.6% and core PCE to 3.3% — both up from 2.7% in March, and both well above the 2% target. A revision that large in a single quarter is the Fed conceding that its disinflation story isn't holding.

And then there's the chair himself. At his first FOMC meeting, Kevin Warsh — a longtime critic of forward guidance — declined to submit a rate projection at all, telling reporters, "I did not submit a dot for me," because it is "not helpful in the conduct of policy." The missing dot is a message: don't read this committee like you read the last one.

He didn't stop there. At the same press conference Warsh announced a slate of internal reviews — task forces on the Fed's communications, the dot plot itself, press conferences, and meeting cadence, alongside a review of the inflation framework. The institution is in motion, not just its rate path.

The deposit book gets uncomfortable fast

Walk into any credit union's deposit committee and look at the CD specials posted this spring. Promotional twelve- and thirteen-month certificates were priced on a curve that assumed the front end was rolling lower by Q4.

If the front end instead rolls higher — even one 25-basis-point hike — every one of those specials is now subsidized funding. The cost of carry against an auto book yielding in the low sevens is a problem you don't fix in a quarter. You either eat the margin compression, let the specials run off without renewal, or pull the offer and absorb the share outflow.

The bigger institutions have desks for this. A mid-size credit union in a secondary market does not. The CFO's spreadsheet just stopped working, and the board meeting is in three weeks.

Auto books were already bending

The other half of the squeeze is the loan side. TransUnion's 2026 consumer credit forecast projects auto loan delinquencies leveling off at an elevated ~1.54% of balances 60+ days past due — a forecast it built explicitly around Fed rate cuts easing borrowing costs. Those cuts haven't come.

Indirect auto, which still drives loan growth for a meaningful chunk of mid-size CUs, is the line item most exposed. Member households that financed vehicles at elevated rates in 2024 and 2025 are now staring at a refi window that may never open. The credit unions that built indirect volume on the assumption that 2026 would bring relief — for themselves and their borrowers — are running a delinquency model that assumes a payment cushion that isn't coming.

Mortgage pipelines built on a fantasy

Real estate teams have been the loudest internal advocates for staffing back up. The argument: rates will drift lower, refi pipelines will reload, purchase activity will follow. Every one of those forecasts referenced the March dot plot.

The June plot says the 10-year doesn't get the help it needed. Mortgage pricing stays where it is, or backs up. Purchase volume stays soft. The new LO you were about to hire is going to underperform her ramp plan.

What an operator does this week

Three concrete moves are worth having on the table by Monday:

  • Re-run the IRR model with a +50 basis point parallel shock as the base case, not the stress case. If the March SEP was your base, you're already behind.
  • Pull the CD special pricing committee meeting forward. Promotional certificate rates that looked smart in May look like a problem now.
  • Stress the indirect auto book against a scenario where the borrower's payment doesn't get any easier for 18 months. Refresh the loss curve. Don't wait for the Q3 ALLL exercise.

The harder Fed dot plot question

The Fed dot plot is a forecast, not a policy. Warsh's missing dot is a reminder that the institution disagrees with itself about where to go next. The June SEP showed a committee genuinely split on 2026 direction — that's a coin flip with member balance sheets on the table.

The credit unions that get hurt most won't be the ones that guessed wrong on direction. They'll be the ones that built a 2026 plan around a single Fed dot plot scenario and never wrote down what happens if that scenario inverts. It just inverted. Now what does your asset-liability committee actually do — and does anyone in the room have the authority to do it before the next dot plot lands in September?