Der Phantom-Verbraucherpreisindex: Blindflug mit gefälschten Inflationsdaten

Key Takeaways

• The 2.4% January CPI is distorted: A 43-day government shutdown erased October 2025 data entirely and contaminated November, forcing the Bureau of Labor Statistics (BLS) to use a “carry-forward” method that creates an artificial downward bias.
• The real number is likely higher: The monthly core CPI of 0.3% annualizes to 3.6%, nearly double the headline year-over-year rate. The carry-forward is masking roughly 0.25-0.3 percentage points of true inflation.
• The contamination lasts until May 2026: Every CPI report from December 2025 through April 2026 is structurally understated. Markets and the Federal Reserve are making decisions on phantom data.
• The Owner’s Equivalent Rent (OER) distortion is the worst: The largest component of the CPI basket has no clean baseline for October, making the entire shelter calculation suspect until fresh survey data accumulates.

The Broken Gauge

On February 13, 2026, the Bureau of Labor Statistics released the Consumer Price Index (CPI) for January. The headline: inflation fell to 2.4% year-over-year, down from 2.7% in December. Core CPI (excluding food and energy) eased to 2.5%, down from 2.6%.

The financial press celebrated. “Inflation cools.” “Progress toward the Fed’s target.” “Rate cuts back on the table.”

There is just one problem: the number is a ghost.

The January CPI is built on a foundation of missing data, statistical estimates, and a technical methodology called “carry-forward” that the BLS was forced to deploy after the longest government shutdown in American history blacked out federal data collection for 43 days last fall. The result is an inflation gauge with a cracked lens, and every institution in the country, from the Federal Reserve to mortgage lenders, is reading it as gospel.

The 43-Day Blackout

The federal government shut down on October 1, 2025, when Congress failed to pass appropriations before the fiscal year deadline. The shutdown lasted until November 12, 2025, spanning 43 days and becoming the longest in U.S. history.

The immediate policy consequences were well-covered. What received far less attention was the quiet catastrophe inside the Bureau of Labor Statistics. Of the BLS’s approximately 2,000 employees, exactly one person worked during the shutdown: the acting commissioner. The rest were furloughed. The agency’s vast network of field economists, who physically survey prices at roughly 23,000 retail establishments and 50,000 housing units every month, went dark.

Here is the timeline of the damage:

Month	Data Collection Status	CPI Report Impact
September 2025	Normal collection	Clean data (last clean month)
October 2025	Zero collection (shutdown)	Report canceled entirely
November 2025	Partial (last 2 weeks only)	Contaminated by holiday pricing
December 2025	Full collection resumed	Distorted by carry-forward baseline
January 2026	Full collection	Distorted by carry-forward baseline
Feb-April 2026	Full collection	Diminishing distortion
May 2026	Full collection	First clean year-over-year comparison

The October CPI report was never published. It is the first time in modern history that a monthly CPI report simply does not exist.

The Carry-Forward Trick

To fill the hole in its data, the BLS employed a technique called “carry-forward methodology.” The concept is simple in principle and dangerous in practice.

When a price collector cannot visit a store or survey a housing unit, the BLS assumes the price in the missing period followed the same trajectory as the prior period. In concrete terms, the September-to-October price change was estimated by repeating the August-to-September price change. The missing October data was then “imputed” based on these carried-forward estimates.

Think of it like a weather station that loses its thermometer for a month. Instead of leaving a gap, the meteorologist writes in last month’s temperatures. If September averaged 72°F, the record will show October at 72°F too, regardless of whether it actually hit 85°F or 60°F. The record looks complete. The record is also wrong.

The Compounding Error

The problem compounds forward. Every year-over-year CPI calculation compares the current month to the same month one year ago. When the October 2025 baseline is an estimate rather than a measurement, the October 2026 year-over-year comparison will also be skewed. More importantly, the direction of the skew is predictable.

Consider the math. September 2025 to October 2025 was the transition from summer to fall, historically a period of rising housing costs (as leases renew and heating costs begin) and rising food prices (as fresh produce becomes scarcer). The carry-forward methodology, by repeating September’s month-over-month change, likely understates the price increases that actually occurred in October.

This means the October baseline is artificially low. And when January 2026 prices are compared to an artificially low October baseline, the year-over-year change appears smaller than it truly is.

The BLS has acknowledged this. Economists at RBC, KPMG, and other institutions have confirmed it. The carry-forward methodology imparts a structural downward bias of roughly 0.25 to 0.3 percentage points on year-over-year CPI readings from December 2025 through April 2026.

The OER Problem: The Biggest Distortion Is in the Biggest Category

The Consumer Price Index is not a single price. It is a weighted basket of roughly 200 categories, and the single largest category is Owner’s Equivalent Rent (OER), a measure of housing costs that accounts for approximately 27% of the total CPI weight.

OER is measured through a survey. BLS economists contact a rotating panel of homeowners and ask, in effect: “How much do you think your home would rent for?” This self-reported estimate is then used as a proxy for the cost of housing in the CPI basket. It is already controversial among economists even in normal times.

During the shutdown, this survey was not conducted. At all.

The carry-forward for OER is especially problematic because housing markets are highly seasonal. October is typically when fall lease renewals push rents higher in many metropolitan areas. By carrying forward September’s OER trajectory, the BLS likely captured none of this seasonal acceleration.

RBC Economics flagged this explicitly, warning that the OER data will not produce a “clean read” until the May 2026 report. They also noted an “exceptionally unusual” divergence: core Personal Consumption Expenditures (PCE), the Fed’s preferred inflation metric that weights housing less heavily, is running hotter than core CPI. Normally, the two track closely. The divergence is a statistical fingerprint of the OER carry-forward dragging CPI artificially low.

$$\text{Apparent CPI}_{YoY} \approx \text{True CPI}_{YoY} - 0.25\text{pp to } 0.3\text{pp (carry-forward bias)}$$

If the reported January core CPI is 2.5% year-over-year, the “true” rate, corrected for the carry-forward, may be closer to 2.8%. And the monthly core rate of 0.3%, annualized, implies underlying inflation of approximately 3.6%, well above what the headline suggests.

The Annualization Gap: What the Monthly Rate Actually Says

Headlines focus on the year-over-year number because it is smooth, stable, and easy to contextualize. But the monthly rate is the raw pulse of price changes, and right now, it tells a different story.

Metric	Reported	Annualized
Headline CPI (monthly)	+0.2%	~2.4%
Core CPI (monthly)	+0.3%	~3.6%
Fed Target	-	2.0%

The 0.3% monthly core rate is not an outlier. It has been the prevailing monthly rate for several months. When annualized, it implies that the underlying pace of inflation is running at nearly twice the Federal Reserve’s stated target.

The year-over-year number of 2.5% looks reassuring only because it is being compared to a contaminated baseline. The monthly number, which is a direct measurement of current price changes, paints a far less comfortable picture.

Who Is Flying Blind?

The Federal Reserve

The Fed’s dual mandate requires it to target 2% inflation. Every FOMC (Federal Open Market Committee) statement references CPI and PCE data as the foundation for rate decisions. If the data is structurally biased downward, the Fed risks making monetary policy on a false signal.

As of February 2026, fed funds futures markets are pricing in two rate cuts starting in June. This expectation is built on the assumption that inflation is genuinely moderating toward 2%. If the carry-forward bias is masking a true rate closer to 2.8-3.0%, those cuts may be delayed or reversed, exactly the “spring surprise” scenario that could rattle bonds.

RBC’s analysis supports this caution, noting that the January data “adds conviction” to the bank’s assumed long Fed pause. The irony is stark: the data looks dovish, but the analysts who understand its flaws are reading it as hawkish.

The Bond Market

The 10-year Treasury yield fell to its lowest level since early December in the days around the CPI release, as investors dumped equities and bought bonds. Long-term Treasury ETFs like TLT rallied 2.8% year-to-date through early February. This flight to safety was partly driven by the AI-related equity selloff, but the “cool” CPI reading reinforced the narrative.

Treasury Inflation-Protected Securities (TIPS) have shown modest gains, with SCHP (Schwab U.S. TIPS ETF) returning +0.57% over the past month. But TIPS payouts are directly indexed to reported CPI. If the reported CPI is artificially low, TIPS holders are being underpaid relative to actual inflation. Their “inflation protection” is being quietly eroded by the carry-forward.

$$\text{TIPS Real Return} = \text{Nominal Return} - \text{True Inflation} \neq \text{Nominal Return} - \text{Reported CPI}$$

Mortgage Borrowers and Savers

Mortgage rates track the 10-year Treasury yield plus a spread. If the “cool” CPI data suppresses Treasury yields, mortgage rates may appear favorable in the short term. But if a spring CPI correction forces yields higher, anyone who locked in a rate in Q1 2026 based on the “inflation is over” narrative may have dodged a bullet, while anyone waiting may face a spike.

On the savings side, inflation-indexed Treasury savings bonds and inflation-adjusted annuities use reported CPI to calculate payouts. A downward-biased CPI means savers are receiving less inflation protection than they are owed.

The Spring Surprise Scenario

The contamination window closes in May 2026, when the first fully clean year-over-year comparison becomes possible. Between now and then, several forces converge to create a potential “spring surprise”:

1. Carry-forward unwind: As real data replaces estimates, the artificial suppression lifts. RBC projects the unwind alone could push core CPI toward 2.8-3.0% year-over-year by Q2 2026.

2. Tariff passthrough: New tariffs imposed in late 2025 are expected to add approximately 0.3 percentage points to core goods prices, with peak impact in Q2 2026, precisely when the carry-forward bias disappears.

3. OER catch-up: The BLS rent survey rotation will begin producing clean OER data around May. If actual rents rose faster in October-November than the carry-forward assumed, the “catch-up” will appear as a sudden jump in shelter inflation.

The convergence of these three forces in the same quarter creates a scenario where CPI could jump from a reported 2.4% to something closer to 3.0% in the span of two or three monthly reports. The headline will read “inflation spikes.” The reality is that inflation never actually fell as much as the data claimed.

This is the 1994 Bond Massacre scenario in miniature: markets price in rate cuts, the data reverses, and the repricing is violent.

The Boring Hypothesis vs. the Cynical One

It is worth stress-testing the cynical read. The boring explanation is that the BLS did its best with an impossible situation. The agency was shut down by Congress, not by choice. Carry-forward is a standard statistical technique. The BLS disclosed the methodology. Economists know about it.

This is largely correct. There is no conspiracy here. The BLS is staffed by serious statisticians who documented the carry-forward and its limitations. The agency did not hide the contamination.

But the boring explanation has a sinister corollary: the political incentive structure around the contaminated data is real, even if the contamination itself is accidental. The White House benefits from lower-looking inflation numbers heading into a midterm cycle. The Fed benefits from data that supports its “wait and see” posture. Mortgage lenders benefit from a rate environment that keeps volume flowing.

Nobody is manipulating the data. But nobody is in a rush to correct the narrative, either.

What To Watch: The May Reckoning

The actionable question is not whether the January CPI is distorted. It is. The question is how much, and when the correction hits.

If the spring CPI reports (March, April, May) show a sudden acceleration, it does not mean inflation is “spiking.” It means the gauge is being recalibrated and the true reading is reasserting itself. The correct frame for interpreting Q2 2026 CPI data is: this is reality returning, not reality worsening.

What to monitor:

• Core PCE vs. Core CPI divergence: If core PCE (which weights housing lower) continues to run hotter than core CPI, the OER carry-forward suppression is still active.
• Monthly CPI rates: Ignore the year-over-year headline until May. The monthly rate is the honest signal. If it stays at 0.3% or higher, the annualized trend is 3.6%, regardless of what the year-over-year number says.
• Fed language shifts: Watch for FOMC statements that acknowledge “data quality considerations” or “base effects.” If the Fed starts qualifying its inflation assessments, it is signaling that it sees the distortion too.
• TIPS breakevens: If the 5-year breakeven inflation rate rises above 2.5% despite “cool” CPI prints, the bond market is pricing in a correction that the headline data has not yet shown.

The Structural Lesson

The 43-day shutdown did not just delay paychecks for federal workers. It broke the measurement infrastructure that the entire economy relies on to make decisions. The CPI is not a luxury statistic. It is the foundation for $7 trillion in inflation-linked contracts, Social Security adjustments, tax bracket calculations, and Federal Reserve policy.

When Congress fails to fund the government, the cost is not just political theater. It is a six-month fog over the economic landscape, during which every major financial institution in the country is, to varying degrees, guessing.

The gauge is cracked. The needle is stuck. And everyone is still reading it.