By Geoffroy Dallennes

Head of Marketing, Exponential Technology Inc.

12/10/2025 at 14:05 New York Time

On the morning of October 10, 2025, the Bureau of Labor Statistics was supposed to release September's Consumer Price Index—the single most important inflation gauge for Federal Reserve policy. The report came out on schedule. Our forecast, published 20 days earlier, predicted headline CPI at +0.304% month-over-month and 3.0% year-over-year. The BLS reported +0.31% and 3.0%. We were off by six-thousandths of a percentage point.

Three weeks later, the government shut down. And it stayed shut down for 43 days.

October's CPI? Canceled forever. October's unemployment report? Lost to history. November's economic data? Delayed until after today's Federal Open Market Committee decision—the most consequential monetary policy meeting of the year.

The Bureau of Labor Statistics went dark. We didn't.

While federal statisticians sat idle, unable to conduct surveys or collect price data, our machine learning models at Exponential Technology continued processing real-time alternative data streams. Energy markets. Transaction flows. Proprietary indicators that don't depend on government operations. We didn't miss a beat.

And what we see in the data the Fed cannot access is deeply troubling: core inflation has been frozen at exactly 3.0% year-over-year for three consecutive months—September, October, and November. Not improving. Not declining. Stuck at 50% above the Federal Reserve's 2% mandate.

Yet today, the Fed is widely expected to cut interest rates for the second time in eight weeks.

This is policy made in the dark. And the consequences could be severe.

When Government Fails, Markets Need Alternatives

I lead marketing at a data analytics firm. My job is to explain complex forecasting models to institutional investors who need to make billion-dollar decisions. I'm not an economist, and I'm not here to tell the Federal Reserve how to do its job.

But I am here to point out something that should alarm every American: the world's most powerful central bank is setting monetary policy for a $28 trillion economy without being able to see the most recent two months of inflation data.

This isn't just about one shutdown or one policy meeting. This is about systemic fragility in America's economic data infrastructure.

For a yera and a half, our team at Exponential Technology—in partnership with LSEG Data & Analytics—has been building AI-powered forecasting systems designed for exactly this scenario. Not to replace government statistics, but to provide redundancy when official data is compromised, delayed, or—as we just witnessed—simply doesn't exist.

Our track record speaks for itself:

• 81% directional accuracy on CPI forecasts 20 days before official releases

• 95% sign accuracy on final forecasts 5 days before BLS reports

• 0.09 basis points mean absolute error over nearly a decade of validated predictions

When we predicted September CPI at 3.0% year-over-year twenty days early, we weren't guessing. When we forecast October core inflation at 3.0% despite the BLS cancellation, we're not speculating. And when we project November core at 3.0%—data the Fed won't officially see until December 18, eight days after today's decision—we're providing the visibility that should be informing policy.

The question is: why isn't the Fed using it?

What the Missing Data Shows

Here's what our models reveal about the inflation picture the Federal Reserve cannot see:

The Three-Month Core Inflation Trend

September 2025 (Actual BLS): Core CPI +3.0% YoY

October 2025 (XTech forecast, BLS canceled): Core CPI +3.0% YoY

November 2025 (XTech forecast, BLS delayed): Core CPI +3.0% YoY

Three months. Zero progress. Core inflation frozen at exactly 3.0%—fifty percent above target.

Meanwhile, headline CPI has been bouncing around based purely on gasoline price swings:

• September: +0.31% MoM (normal gas prices)

• October: +0.2% MoM (gas fell, creating false disinflation signal)

• November: +0.33% MoM (gas surged +3.1%, reversing the decline)

Here's the dangerous part: The Fed doesn't know what happened in October. Markets don't know what happened in October. With September showing inflation at 3.0%—an acceleration from August's 2.5%—and no official data for the next two months, policymakers are making decisions based on hopes and assumptions rather than actual evidence.

The assumption? That inflation is cooling, that the September acceleration was noise, that the path back to 2% is proceeding smoothly.

Our data says otherwise. Core inflation—the measure that strips out volatile energy and food prices—never moved. It stayed at 3.0% in October and November while gasoline volatility created headline swings that no one but us can see.

The Fed has historically emphasized core inflation precisely because headline can be distorted by supply shocks the central bank can't control. OPEC production decisions. Refinery disruptions. Seasonal demand patterns. These aren't demand-driven inflation that interest rate policy can address.

But in the fog of the government shutdown, with official October data lost forever and November delayed past the FOMC meeting, the Fed is flying blind—making assumptions about disinflation that our real-time data contradicts.

The Shelter Problem No One Wants to Talk About

Dig deeper into our category-level forecasts, and the core inflation persistence becomes even more concerning.

Shelter costs—representing 35% of the entire CPI basket—rose +0.3% month-over-month in both October and November. That translates to 3.6% annualized inflation, nearly double the Fed's 2% target.

Do the math: With shelter at 3.6% annualized and 35% weighting, you need near-deflation in other categories just to average 2% overall. That's not happening. Used cars are flat (normalization complete). Food inflation continues at ~2.5%. Services remain sticky around 3%.

The only path to 2% inflation is sustained restrictive policy to cool housing demand. But the Fed already cut rates by 25 basis points in October. And today, they're considering cutting again—to 3.50-3.75%, which would leave real interest rates (adjusted for core inflation) at just 0.5-0.7%.

That's barely restrictive. That's accommodation while shelter inflation runs at 3.6% and core sits at 3.0%.

Academic research on monetary transmission mechanisms is unequivocal: interest rate changes take an average of 29 months to fully affect price levels. The cut the Fed makes today influences inflation in Q2 2028. The October cut won't even begin showing up in CPI data until Q3 2026.

You're easing before you've seen the impact of the last cut. That's not data-dependent policy. That's hope.

The 1970s Parallel No One Wants to Hear

I wasn't alive during the Great Inflation. But I've read the academic research on how the Federal Reserve lost control of prices in the 1970s, and the parallels are uncomfortable.

The definitive NBER analysis is blunt: the fundamental error was that "short-term interest rates responded too little and too late to rising inflation." Policymakers repeatedly cut rates during brief disinflationary episodes, hoping each dip signaled the start of sustained price stability. It never was. And premature easing made the eventual Volcker shock both necessary and brutally painful.

Three specific mistakes from that era are visible in today's environment:

1. Responding to headline inflation driven by supply shocks (oil embargoes then, gasoline volatility now) rather than focusing on core

2. Cutting rates prematurely when inflation showed brief moderation, before structural pressures normalized

3. Prioritizing employment concerns over inflation control, leading to unanchored expectations

Research on successful disinflations finds three common factors: "clear-cut goals, a willingness to accept higher unemployment, and a refusal to give up prematurely."

Paul Volcker's success in the early 1980s came from that third factor. In 1981, after initial easing in late 1980, the Fed reversed course and tightened aggressively—pushing the federal funds rate into the 16-22% range—despite recession. That sustained commitment, maintained through economic pain, broke inflation expectations and restored credibility.

The lesson isn't about specific rate levels. It's about resolve.

Cutting rates twice in eight weeks while core inflation sits frozen at 3.0% doesn't demonstrate resolve. It demonstrates capitulation to market expectations and political pressure.

Why Alternative Data Can't Stay Alternative

The 43-day shutdown exposed a vulnerability that extends far beyond one policy meeting: America's economic policy infrastructure is a single point of failure.

When BLS operations stop, the entire edifice of data-driven policymaking collapses. Not just monetary policy—fiscal policy loses visibility into tax receipts. Businesses can't plan investment without economic indicators. Financial markets experience elevated volatility from information gaps. International institutions lose confidence in U.S. data quality.

This is unacceptable for the world's reserve currency issuer.

Our Macro Predictions platform demonstrates that private-sector forecasting can provide reliable, early inflation intelligence that operates independently of federal survey infrastructure. We're not proposing to replace the BLS—nothing should replace the gold-standard comprehensive methodology of official government statistics.

But official statistics that arrive late, incomplete, or not at all cannot be the only input to the most consequential policy decisions on Earth.

The Federal Reserve should formally integrate alternative data sources into FOMC materials. Not as replacements, but as redundancy. As validation. As insurance against exactly what just happened.

Consider what alternative data provides:

• Operational independence: Forecasts continue when government operations cease

• Timing advantage: Intelligence available 20+ days before official releases

• Granularity: Category-level visibility distinguishes energy noise from core trends

• Validation: Eight years of demonstrated accuracy (September: 0.006pp error)

• Real-time visibility: When official data disappears, alternative data fills the void

The November CPI report arrives December 18—eight days after today's FOMC decision. By then, rates will have been set. Markets will have reacted. The window for data-driven policymaking will have closed.

Alternative data ensures that window never has to close.

What the Fed Should Do Today

I'm not naive. I understand the institutional and political pressures the Federal Reserve faces. Markets price an 83-88% probability of a rate cut today. The forward guidance has been clear. The pattern of sequential cuts, once easing begins, is well-established.

But central bank independence means exactly this: the courage to follow economic reality rather than market expectations.

The data we can see—and the Fed cannot officially access—says core inflation has made zero progress in three months. Shelter inflation persists at 3.6% annualized. Month-over-month core runs consistently at +0.3% (3.6% annualized). Real interest rates are already barely restrictive at ~0.9%.

The right call is to hold rates at 3.75-4.00%.

Signal that you're assessing the impact of October's rate cut before making sequential moves. Make clear you distinguish between headline energy volatility and core inflation persistence. Emphasize that core at 3.0% for three straight months isn't progress—it's stagnation at an unacceptably elevated level.

If political realities make a hold impossible, then at minimum:

• Make this the final cut of the cycle with explicit, quantitative guidance

• Acknowledge that core inflation has not improved despite October's apparent headline disinflation

• Show in the dot plot that no further cuts are expected in 2026

• Define precisely what "sustained progress" means: three months of core at or below 0.17% MoM

But the cleanest, most credible path is simply: don't cut twice in the dark.

The Broader Stakes

U.S. monetary policy doesn't just affect Americans. Academic research confirms that Fed rate decisions transmit rapidly to emerging market interest rates through capital flows and exchange rate dynamics—even in countries with flexible currency regimes and inflation targeting.

Your rate cut today reverberates through over $60 trillion in dollar-denominated global debt. Borrowing costs, refinancing ability, and debt sustainability for dozens of nations depend on whether the Federal Reserve maintains credible commitment to price stability.

When the Fed eases prematurely while core inflation remains frozen at 3.0%, you're not just taking risks with American price stability. You're forcing risk onto the entire dollar-based international monetary system.

The U.S. dollar's reserve currency status is a privilege that carries responsibility. Part of that responsibility is maintaining the highest standards of data-driven decision-making—even when, especially when, official data is compromised.

A Call to Action

To the Federal Reserve: Use the data that exists. Alternative forecasts from validated private-sector models should inform FOMC deliberations when official statistics are delayed or missing. Our September forecast was accurate to 0.006 percentage points. Our October and November forecasts show core inflation frozen at 3.0%. This intelligence is available. Use it.

To Congress: Protect economic data infrastructure as critical national infrastructure. Government shutdowns should not cancel CPI releases or employment reports. Fund the Bureau of Labor Statistics and Census Bureau with mandatory appropriations that cannot be interrupted by political dysfunction. Make it illegal to defund statistical agencies during shutdowns.

To the broader policy community: Take alternative data seriously. The government shutdown proved that official statistics can disappear when you need them most. Private-sector forecasting provides redundancy, validation, and insurance. It's time to integrate it formally into the policy process.

To investors and markets: Look at core, not headlines. Gasoline volatility is noise. Core inflation frozen at 3.0% for three months is the signal. Don't mistake energy-driven headline swings for genuine disinflation.

The Data Exists. It Shows Core Frozen at 3%. Will Anyone Listen?

Twenty days before the October 10 BLS release, we predicted September CPI at 3.0% year-over-year. We were exactly right. Today, we're telling you that October and November core inflation remained at 3.0%—unchanged, unimproved, stuck fifty percent above target.

The Federal Reserve won't officially see this data until December 18. By then, today's decision will be history.

We've spent eight years building the systems to see through the fog. The instruments exist. The visibility exists. The data exists.

The question is whether the most powerful central bank in the world will make policy based on validated alternative intelligence, or whether it will cut rates twice in eight weeks while operating blind—hoping that gasoline price swings signal disinflation that isn't happening.

Core inflation at 3.0% for three straight months isn't progress. It's policy failure in waiting.

Don't cut twice in the dark.

Geoffroy Dallennes is Head of Marketing at Exponential Technology Inc. We empower enterprises to innovate more rapidly by leveraging foundational data technology in combination with tightly integrated analytics and LLM toolsets. By federating data into a high-performance and flexible hybrid cloud backplane we help businesses of all sizes overcome data entropy and unlock their true analytical capacity as an organization The views expressed are his own.

Contact: sales@exponential-tech.ai 

| +1 (807) 703-3188 |

exponential-tech.ai

Exponential Technology's Macro Predictions platform delivers CPI and PCE forecasts 20 days before official BLS releases, with 81% directional accuracy, 95% sign accuracy, and 0.09bp mean absolute error validated over eight years of continuous forecasting since November 2017.