The Board Turned Greener While the Fed Turned Hawkish

Observation dates for the indicators above range from April 2026 (delinquency rate) through September 29, 2026 (VIX, Bitcoin, and live yield curve readings), with most series at August 2026. Each figure is described in this post by the date it was observed, not by the month of publication.

Pearl Quest EconPulse for September 2026 shows a signal board that improved in two places: growth returned to Supportive as the third quarter matured, and consumer confidence moved off red for the first time since spring. In the same month, the Federal Reserve raised interest rates, the yield curve flattened by twenty basis points, and consumer sentiment fell to one of the lowest readings on record. The board and the world moved in opposite directions.

Last month the seven signals held perfectly still while the ground moved underneath them. This month the board finally moved, and it moved the wrong way.

Two signals improved. Growth went green as the quarter matured past the point where our nowcast caution applies. Consumer confidence came off red for the first time since spring. Read the board alone and September looks like the month conditions turned.

September was the month the Federal Reserve raised interest rates for the first time in this cycle, the month the yield curve flattened hard, and the month consumer sentiment fell to 48.1, within sight of its all-time low. None of those three events appears anywhere on the board above.

That gap is the story, and it is no longer a curiosity. It is a framework problem we have to name plainly.

The Fed Hiked

On September 16 the FOMC raised the federal funds target range by 25 basis points to 3.75% to 4.00%. The vote was unanimous. Chair Warsh was blunt about why: "The plain fact is that inflation is too high and has been for too long." The Committee's own projections showed twelve of eighteen members expecting one more increase before the end of the year.

The effective federal funds rate stood at 3.88% as of September 28, up from 3.63% through the summer. This is the first increase of the cycle, and it arrives after nine months of a rate that had not moved.

Read the correction carefully, because the interesting part is not that our number was stale. It is which number we chose to read.

EconPulse already reads the yield curve daily. Both spread rows in this month's brief are marked "Live." But it reads monetary policy from the monthly average, a series that by construction cannot show a mid-month policy change until the following month. We picked a lagging series for the one variable in the framework that moves on announced dates, and then described a tightening cycle as an easing one.

Moving these briefs to the end of the month, which we are doing for other good reasons, would not have caught this. The monthly average for September will not publish until October regardless of when we generate the brief. The fix is to read the daily rate, exactly as we already do for the curve.

We flagged a version of this in August, when we wrote that the framework "has no market-implied policy-path input" and was "scoring an easing cycle while the market prices tightening." The market was right, and the gap turned out to be simpler and more embarrassing than a missing forward-looking input. We were reading the wrong series for the rate that already exists.

The Curve Flattened, Which Is Not What the Board Says

The yield curve row reads Supportive, and on the narrow question of whether the curve is inverted, that is correct. It is not. But the shape changed sharply this month, and the direction is the opposite of last month's.

As of September 28, the ten-year Treasury yielded 5.24% and the two-year 4.92%, a spread of 32 basis points. In August that spread was 52 basis points. The curve flattened by twenty basis points in a single month.

The mechanism matters. A month ago the curve was steepening because long yields were climbing, which we described as the bond market demanding a larger term premium for inflation risk. This month the short end caught up and passed it. The two-year yield is responding to a Fed that just hiked and is signaling another. Long yields rose too, with the ten-year now above 5.2%, but the front end rose faster.

A flattening curve driven by rising short rates is the classic shape of a tightening cycle in progress. It is early, the curve is still positively sloped, and a positive slope is genuinely better than an inverted one. But "Supportive" and "flattening toward inversion under an active hiking cycle" are describing the same number and telling different stories.

Inflation: Cooling in the Aggregate, Burning at the Pump

The framework carries CPI at 3.7% year over year as of August and scores it Neutral. The Bureau of Labor Statistics reported 3.4% for the same month, released September 11. Our framework runs consistently high by roughly a quarter of a point, a discrepancy we have been tracking since July and most likely a seasonal-adjustment mismatch. A reader checking BLS directly will see the lower number, and the lower number is the official one.

Either way, the headline cooled. Core inflation, which strips out food and energy, came in at 2.4% as of August, its best reading in this cycle.

The composition is where the trouble sits. Energy prices rose 16.3% year over year as of August. Gasoline rose 27.4%, and gasoline alone accounted for more than a third of the monthly increase in the all-items index. Core is behaving. The part of the index that households buy weekly, cannot defer, and see advertised in foot-high numbers on the way to work is not.

This is the mechanism that connects everything else in this post. It is why the Fed hiked against a cooling headline number. It is why sentiment collapsed while core inflation improved. And it comes from a market our framework does not yet track.

Oil, Outside the Framework

Brent crude futures stood at $97.82 a barrel and WTI at $90.27 as of September 30, up roughly 8% and 5% over the past month and about 50% and 46% from a year ago.

The path through 2026 has been violent. Brent averaged $70.89 in February, jumped to $103.13 in March, and peaked at a monthly average of $117.29 in April. It fell back to roughly $84 in June and July as the immediate conflict premium drained out, then climbed again to $91.08 in August and higher since.

Middle East crude exports have recovered to approximately 98% of pre-war levels, with Saudi Arabia restarting shipments through its East-West pipeline, but flows remain below pre-conflict levels and the market is still characterized as undersupplied. Washington has signaled no intention of easing sanctions on Iran, and OPEC+ is expected to hold November quotas unchanged.

Energy is the transmission channel running through this entire brief: to the CPI print through gasoline, to the Fed's decision through inflation expectations, and to household sentiment through the price at the pump. EconPulse has no oil indicator, which means the single most important macro variable of 2026 enters our framework only indirectly, months later, through CPI.

We are adding crude oil to snapshot capture. Under our standing policy, a new indicator is captured immediately and presented only once enough history has accrued for its z-score to mean something, so it will appear in the table in a future brief rather than this one.

Growth Returns to Green, With a Caveat About What It Measures

GDPNow reads 3.7% for the third quarter, with a z-score of +0.65 and an 85th-percentile standing against its five-year history.

Last month we capped this signal at Neutral under our Early-Quarter GDPNow Exception, which applies while fewer than two of a quarter's three months have complete source data. The third quarter is now essentially complete, so the cap no longer applies and the signal stands at its native Supportive.

Worth remembering how this quarter behaved. Q3 opened at 5.0% on July 30, spiked to 6.2% by August 3, and then fell in every subsequent update: 5.8%, 4.3%, 4.0%, and now 3.7%. The nowcast has lost roughly two-fifths of its initial estimate over the quarter, with the revisions concentrated in private investment. A 3.7% quarter would still be strong. The trajectory is a steady walk downward from a number that was never reliable.

Q2 finished at 1.5%. If Q3 lands near 3.7%, the first half of 2026 will have been considerably weaker than the second.

Labor and Credit Still Hold

Unemployment sits at 4.1% as of August, unchanged from July, with a z-score of -1.73 against its one-year history. The labor market remains the strongest single argument against a downturn narrative, and it has now held near these levels for three months.

Credit conditions are benign. The default spread stands at 1.64% as of August, up modestly from 1.59% in July and sitting at the 28th percentile of its five-year range. Wider, but comfortably within normal territory.

The consumer delinquency rate returns to the board this month. We suspended it in August under our frozen-indicator rule after it sat unchanged at 2.64% with a January observation date for four consecutive snapshots. Fresh data has resumed: the rate now reads 2.62% as of April 2026. It is still four months stale, which is inherent to the series, but it is moving again and it is informative again. Credit Conditions no longer rests on the default spread alone.

Consumer Confidence Came Off Red at Exactly the Wrong Moment

The framework carries consumer sentiment at 52, the August observation, at the 12th percentile of its five-year range. That is an improvement from the 3rd percentile reading it carried all summer, and it is enough to move the signal from Concerning to Neutral.

This is the most misleading number on the board this month, and it is worth being precise about why.

The University of Michigan released its final September reading on September 25, five days before this brief was generated. Sentiment fell to 48.1, down 7% from August. The expectations component fell 10.4% to 46.3. Year-ahead inflation expectations jumped to 4.6% from 4.0%, the highest since June. Survey director Joanne Hsu attributed the decline to concerns about high prices continuing to climb and to worries that elevated fuel prices could pass through to the broader economy.

So the board upgraded consumer confidence from red to yellow in the same month that actual consumer confidence fell to within a few points of the lowest readings ever recorded in the series.

The mechanism is publication lag, which we have documented at length and which is not a bug. Michigan's series reaches our data source with a one-month delay, posted around the 25th of each month, so this brief describes August while households are living in late September. The September figure enters the framework with the October 23 update, which means our next brief will carry it. But this month the lag did not merely make the signal stale. It made the signal point the wrong direction. An indicator that improves on the board while deteriorating in reality is worse than an indicator that is simply late.

That timing is also why these briefs are moving to the end of the month. Capturing after the sentiment update, rather than before it, makes every indicator on the board a uniform one month old instead of leaving this one series an extra month behind the rest.

Treat the yellow light on the Consumer Confidence row as an artifact of timing. The September reading is the one that describes where households actually are.

Hedges and Risk Appetite

Gold is up 21.2% year over year as of August, roughly flat against last month's 20.1% pace, and continues to read Supportive as an inflation hedge. Bitcoin is up 10% year over year as of September 29 and reads Neutral.

VIX at 16.0 as of September 29 is essentially unchanged from August's 15.8 and remains in normal range. Equity volatility markets are not pricing stress, which is its own kind of information given a rate hike, a flattening curve, and near-record-low consumer sentiment all landing in the same month.

What the Data Doesn't Show Yet

Three items sit outside the board this month.

What comes after the hike. The increase itself is corrected in the body of this post. What sits outside the framework entirely is the guidance: twelve of eighteen FOMC participants project another increase before year-end, and Chair Warsh framed the September decision around a "timelier return" to the 2% target. The framework scores the level of the policy rate and has no input for its expected path, so a second hike will register only after it happens.

September sentiment at 48.1. Discussed above, and the single most important number in this post that does not appear in the table. The expectations component at 46.3 and year-ahead inflation expectations at 4.6% both point the same direction. Households are not responding to the improving core inflation print. They are responding to the price of gasoline and to a forward view that has deteriorated across the political spectrum. This reading enters the framework on October 23 and will appear on next month's board.

Energy as the common cause. Gasoline up 27.4% year over year, energy up 16.3%, Brent futures near $98, and a supply picture still constrained by the Gulf situation. This one variable plausibly explains the Fed's decision, the sentiment collapse, and the composition of the CPI print. It is the largest thing the framework cannot see, and it is the reason we are adding crude oil to capture this month.

What to Make of All This

The September board reads better than August's. Two signals improved and none deteriorated. That is the opposite of what happened.

• The Fed hiked for the first time this cycle, to 3.75%-4.00%, with another increase projected by two-thirds of the Committee.
• The yield curve flattened by 20 basis points, the front end responding to a tightening Fed, while the signal stayed green.
• Consumer confidence improved on the board and collapsed in reality, falling to 48.1 in a September reading the framework will not carry until next month.
• Core inflation reached its best level of the cycle at 2.4%, while gasoline rose 27.4% and drove a third of the monthly CPI increase.
• Growth returned to Supportive, on a nowcast that has been revised down steadily all quarter.

A note on what this meant for small caps. The third quarter was unusually hard on smaller companies, and the reasons turn out not to be the obvious ones. We took that apart separately in The Size Factor Measured Three Different Things in 2026, which finds that the familiar "small caps versus the S&P" spread was measuring three different things across the three quarters of this year, and that what looked like a small-cap event in the third quarter was substantially a story about the ten largest companies in the index. That note ships with the data and a script so any reader can check its arithmetic.

There is a pattern in these five items, and it is not flattering to our framework. Every one of them is a case where a backward-looking measure of a variable improved while the forward-looking version of the same variable deteriorated. That is the defining feature of a turning point, and it is precisely the condition under which a framework built on realized data performs worst.

We said last month that realized and expected inflation had separated. This month the separation widened and spread to monetary policy, to the curve, and to households. The board is not wrong about the past. It is measuring a past that the present has already left behind.

What to watch: whether the October CPI print holds core near 2.4% or absorbs the energy pass-through, whether the Fed delivers the second hike its projections imply, whether the curve keeps flattening, and whether the October sentiment reading confirms September's collapse or rebounds. If the curve inverts while the Fed is still hiking, this framework will finally register what the bond market has been saying since August.

About EconPulse

EconPulse is a Pearl Quest macro monitoring framework that evaluates key indicators of economic momentum, financial conditions, and market stress to classify the current economic regime. You can see the underlying charts and generate your own brief anytime here: https://econpulse.pearl-quest.com/