WASHINGTON, 7 November 2025 — With official U.S. government data sources frozen amid the prolonged federal shutdown, the U.S. Federal Reserve (Fed) has turned to private data providers, artificial intelligence (AI) tools, and business surveys to fill the information void ahead of its December policy meeting.
Two weeks before the Fed’s previous meeting on October 28–29, the Atlanta Federal Reserve relied on alternative datasets, ranging from online job postings to corporate sentiment analysis, to gauge the economy’s health. Staff analysts studied how past business surveys aligned with employment, spending, and output trends, boosting their confidence in these substitutes despite the lack of official statistics.
“We are getting a pretty solid read on what is happening in aggregate even if we don’t have the official statistics,” said Brent Meyer, Assistant Vice President and Head of the Atlanta Fed’s Economic Survey Research Center. “This isn’t doing things by anecdote. We are capturing a large swath of what is happening.”
Private Data Becomes Policy Lifeline
Fed Chair Jerome Powell (NASDAQ:GOOGL) previously cautioned that the data hiatus could make policymakers more measured about future rate decisions. However, the central bank now finds itself “drowning in information,” according to staff, as it blends AI-driven analytics, online data scraping, and cellphone mobility tracking to identify economic signals from vast new datasets.
Some of these methods have proven reliable. Online job postings from firms like Indeed have tracked closely with official labor reports. But other areas, such as services inflation, remain opaque. Efforts are underway to train AI models to analyze company earnings reports for pricing trends.
Ahead of the October meeting, the Atlanta Fed concluded that third-quarter activity had “ticked up” and that the economy was holding steady, even as price pressures persisted. Investors now expect another quarter-point rate cut when the Fed meets on December 9–10.
Fed’s Expanding Toolkit: From Beige Book to AI Models
Across the central bank system, officials have launched new data dashboards, modeled unemployment estimates, and sentiment analyses of the Beige Book, the Fed’s qualitative economic report, to compensate for missing federal releases.
A Boston Fed study is using natural language processing to flag potential recession risks based on Beige Book wording trends. Meanwhile, the Chicago Fed and private firms such as Revelio Labs and ADP are providing real-time estimates for job creation, unemployment, and wage trends.
“Key data points that investors rely on to give them some kind of sense of direction … are just not available,” said Meyer. “But we are finding that many of these private datasets track overall growth really closely.”
The shutdown has delayed several critical indicators, including the non-farm payrolls report, the Consumer Price Index (CPI), and unemployment data, a first in modern U.S. history.
Limitations and Human Insights
Fed officials acknowledge that while private datasets provide speed, they lack the demographic and geographic depth of government statistics.
For instance, Revelio Labs, a seven-year-old analytics firm scraping employment data from platforms like LinkedIn, can estimate monthly job changes but not broader metrics such as labor force participation.
Similarly, Harvard professor Alberto Cavallo, founder of PriceStats, said his AI-powered price database can track goods inflation but lacks coverage for housing and services.
“The data is meant to be directional. We can’t replace the CPI,” Cavallo said during a Cleveland Fed webinar.
Fed Governor Lisa Cook echoed that a blend of high-frequency data and direct human feedback remains crucial. “We are looking at a panoply of data in real time,” she said. “But what I want is the mortar between the bricks, to hear what the stories are.”
Source: Reuters
