Economy March 6, 2026

Private-sector data could sharpen Fed decisions, researchers say

Study finds combining payroll processor, brokerage and bank data with official reports would have improved job and inflation forecasts during a turbulent year

By Derek Hwang
Private-sector data could sharpen Fed decisions, researchers say

A group of economists found that melding anonymized private data from payroll processor ADP, Vanguard 401(k) activity and JPMorgan checking accounts with government employment and inflation reports produces more accurate forecasts of monthly private nonfarm payrolls, subsequent revisions, and six-month job growth. The team says private signals - especially on smaller firms and lower-income households - could have helped Federal Reserve officials anticipate labor-market weakening and inflation trends amid the disruptions of a prolonged government shutdown and large revisions to jobs data.

Key Points

  • Combining anonymized ADP, Vanguard 401(k) and JPMorgan checking data with official reports improved forecasts of monthly private nonfarm payrolls, two-month revisions, and six-month-ahead job growth.
  • Private income and wage signals from anonymized data aided in anticipating consumer inflation readings.
  • Private data appear especially informative for smaller firms and lower-income households, areas that can be underrepresented in government surveys - this has implications for labor, consumer-facing sectors and financial markets tied to interest-rate expectations.

A new study of anonymized private-sector indicators concludes that augmenting government economic reports with selected private data would likely improve the Federal Reserve's ability to track jobs and inflation in near real time. The researchers say a blended data approach would have offered clearer signals during a recent year in which official statistics were interrupted by a record-long federal government shutdown and then clouded by unusually large revisions.

The team examined how series from payroll processor ADP, anonymized 401(k) information from Vanguard, and aggregated checking account activity from JPMorgan could complement the monthly employment and consumer-price reports produced by government agencies. They report that combining public and private inputs improved forecasts of the change in private nonfarm payrolls published by the Bureau of Labor Statistics each month, enhanced predictions of the subsequent two-month revisions to that number, and strengthened six-month-ahead job growth forecasts.

Because the anonymized private records also contained income and wage information, the researchers found those series helped anticipate consumer inflation readings as well. The work was presented at a Booth monetary policy conference in New York and was produced by economists from the three organizations that contributed data, together with University of California at Berkeley professor Yuriy Gorodnichenko and University of Chicago Booth School of Business professor Anil Kashyap.

The researchers underscore that officials already rely on private inputs in routine analysis - credit card spending and a variety of bank and industry reports have long been part of the Fed's toolkit - but they say interest in alternative, higher-frequency sources strengthened during the pandemic. The health crisis produced volatile economic swings and contributed to a decline in survey response rates, prompting a search for additional indicators that move closer to real time.

The research team highlighted how private data could have clarified a pivotal episode in the Fed's policy debate last year. ADP payroll data, they report, signaled that labor-market momentum was weakening beginning that spring, even as initial Bureau of Labor Statistics reports for May and June continued to show strong job gains. Those early official readings were later erased in subsequent revisions, a sequence of events that the researchers say helped prompt the Fed to proceed with interest-rate cuts in September and coincided with the firing of the BLS commissioner by President Donald Trump.

Federal Reserve Governor Christopher Waller is cited in the study as having warned about likely revisions while referencing ADP data, urging colleagues to move forward with rate reductions in the summer and dissenting at the July Federal Open Market Committee meeting when the committee voted to hold rates steady.

The study's authors present a pragmatic view of private data's role. They note that a hybrid approach is designed to support official statistics rather than replace them. Still, private records appear to contribute information in segments of the economy that government surveys may capture less well, notably smaller firms and lower-income households - groups that can be underrepresented when sample sizes and response rates are constrained.

The authors write, "Signals from...private data could potentially reduce the noise that policymakers face," and add that, "Private sector data contribute a signal precisely in the areas of the economy, such as smaller firms and lower-income workers, where government surveys - constrained by sample sizes and response rates - may be weaker." They also suggest that incorporating an even broader set of private sources than those used in this study could further strengthen predictive performance.

One operational challenge the paper emphasizes is the sheer volume and heterogeneity of privately available information. Modern data streams span cellphone location traces, point-of-sale barcode scans and myriad other high-frequency records. Identifying series within that mass of raw data that consistently add forecast value, possess a sufficiently long track record for study, and can be benchmarked against eventually revised government statistics is nontrivial.

The research also reiterates a distinctive feature of official employment statistics: the ‘‘real‘‘ payroll number is effectively a moving target that takes more than a year to settle. Initial government surveys are revised twice as additional responses come in from businesses, and later the survey-based series is benchmarked to a broader quarterly census of business activity that nearly all firms complete. Policymakers, however, cannot wait a year for that final figure and must monitor economic developments close to real time in order to react when conditions shift.

To address those constraints, the team constructed an index from the private sources and found it improved short-term and medium-term forecasts relative to relying on government releases alone. That index, they report, showed earlier signs of labor-market softness in the spring that the initial BLS releases did not capture at the time.

The authors are careful to frame private data as an enhancement to public statistics rather than a substitute. They argue that the complementary strengths of both types of data - the representativeness and methodological rigor of government series and the timeliness and granularity of some private indicators - make a combined approach promising for monetary policymaking.

While the study focuses on specific private contributors and a set of employment and inflation outcomes, the researchers note that broader adoption of vetted private sources could yield further gains in real-time economic assessment. They caution, however, that isolating useful signals from a growing universe of digital traces will remain an analytical hurdle for policymakers seeking to reduce uncertainty without sacrificing the methodological integrity that underpins official statistics.


Summary

Researchers analyzing anonymized ADP payroll data, Vanguard 401(k) flows and JPMorgan checking account activity find that combining those private indicators with government employment and inflation reports produced more accurate forecasts of monthly private nonfarm payrolls, later revisions, and six-month job growth. The blended approach also aided in anticipating consumer inflation, and would have provided earlier signals of labor-market weakening than initial official releases during a year marked by a federal government shutdown and large BLS revisions.

Risks

  • Large volumes of private-sector raw data make it difficult to identify series that consistently add predictive value - this analytical challenge could limit reliable adoption and affects sectors reliant on timely labor and consumer metrics, including financial markets and retail.
  • Government employment statistics are revised over a long period - the ‘‘real‘‘ payroll number takes over a year to settle - meaning policymakers must balance early private signals against final official benchmarks, a tension relevant to monetary policy decisions and interest-rate sensitive sectors.
  • Private data can supplement but not replace official reports; overreliance on limited private sources could mislead if those sources do not represent the broader economy, posing risks to macro-sensitive industries and central-bank decision making.

More from Economy

FDA Vaccines and Biologics Chief Vinay Prasad to Exit Agency at Month-End Mar 6, 2026 Where Fed Officials Stand: A Snapshot of Policy Views Ahead of the March Meeting Mar 6, 2026 White House Secures Deal to Quadruple Output of 'Exquisite Class' Arms Mar 6, 2026 Cleveland Fed President Calls February Jobs Report a Disappointment, Urges Caution on Policy Mar 6, 2026 Pimco Warns Direct-Lending Could Face Its First Major Default Cycle as Investors Seek Redemptions Mar 6, 2026