What happens when US economic data becomes unreliable

Capturing the complexity of the U.S. economy is a formidable task. Accurate data collection involves millions of individuals gathering and sharing data across millions of establishments, resulting in billions of decisions based on that data once it’s been aggregated. To meet this challenge, the U.S. relies on 13 major statistical agencies that provide important data on labor, health, economics, education, and agriculture.

Yet recent political interference, shrinking agency budgets, and low response rates to government data surveys have created ruptures in the system and led to a growing public mistrust of institutions. 

There are numerous consequences to having unreliable data, said MIT Sloan professor of applied economics Roberto Rigobon, a research associate of the National Bureau of Economic Research. Among them:

• Policymakers may misjudge the economy’s health.
• Investors may lose confidence in the reliability of the data.
• The public may disengage from participating in official measures altogether.

In a working paper titled “Measuring by Executive Order,” Rigobon and Harvard Business School professor Alberto Cavallo address the main challenges undermining trustworthy government data and detail what businesses should be aware of, especially regarding the use of private data. 

3 current challenges with U.S. data 

1. Declining survey response rates. Statistical agencies depend on routine surveys of households and companies to construct measures of employment, inflation, and other core indicators, but response rates have fallen dramatically in recent decades. In the past, people were more willing to answer surveys over the phone or in person, but that’s changing. 

   “People have stopped answering the phone,” Rigobon said. This is a problem because low response rates introduce bias, delay revisions, and weaken the representativeness of key statistics.

2. Funding constraints. Government agencies like the Bureau of Labor Statistics and Census Bureau are facing shrinking budgets that limit their ability to adopt new technologies and expand data-collection efforts. One example: In September 2025, the U.S. Department of Agriculture announced that it was halting its “costly” annual survey on food insecurity, which will prevent policymakers and researchers from tracking changes to household hunger in the U.S.

   “It has become really, really difficult for the statistical offices to collect the data points,” Rigobon said. “Why this is so important? Because you need representativeness. Representativeness is by far the most important attribute of accurate data.”

3. Political interference. Breaking apart advisory committees, dismissing statistical leaders, and politicizing nominations may not immediately alter data quality, the authors write, but those actions undermine transparency and credibility. Countrywide government shutdowns, which include statistical offices, have far-reaching ramifications. 

   “The shutdowns that happen, they tend to be really costly for the statistical offices because they cannot collect the data,” Rigobon said. Losing one month’s worth of data is considerable when you have only 12 months’ worth of data to begin with. “One data point is a lot,” he said.

   Likewise, revisions to U.S. government data are routine; agencies make scheduled updates to initial, often preliminary, statistical estimates to enhance accuracy. But lately those revisions have come under attack, with some people characterizing them as a sign of failure or bias.

   “Policymakers may rely on preliminary numbers to act quickly, while investors and analysts turn to revised data for a clearer long-term picture,” the authors write. “Far from signaling failure, revisions are a hallmark of a healthy statistical system that adapts as better information becomes available.”

Future-Ready Enterprise Academy

In person at MIT Sloan

Register here

Action items for business leaders

1. Use private data, but with caution. Private-sector data can play a useful role in complementing government data, especially as survey response rates decline. 

Whether collected by academics, financial institutions, or technology firms, private data is useful as an independent source that can provide a check on official numbers, highlight discrepancies when they arise, and fill in the blanks where government data falls short. 

However, private-sector data cannot fully replace official statistics for a number of reasons, including: 

• Coverage. Private-sector data cannot match the breadth of official surveys, especially for complex measures such as employment, inequality, or production in small firms and local markets.
• Incentives. Because private data is often produced to meet commercial demand, areas with broad social value may be neglected.
• Transparency. Many providers rely on proprietary methodologies that are rarely disclosed in detail, limiting transparency and making replication difficult.

In short, “a healthy economy benefits from a robust interplay between official and private statistics, each reinforcing the other’s credibility and value,” the authors write. 

2. Speak up. The integrity of economic data is an important component of democratic governance and market stability, Rigobon said. Vigilance is essential for detecting and resisting political manipulation in its early forms before public trust slips away and becomes difficult to regain. 

To that end, companies should be speaking up more. “It’s time for them to stand up and say, ‘These policies make no sense,’” Rigobon said. Specifically, companies aren’t fully grasping the implications of staying silent on tariffs. “It’s a tax on firms, and firms should be more vocal,” he said. 

Ultimately, reliable statistics require investment, institutional independence, and public trust, the authors conclude. “Protecting and strengthening the U.S. statistical system is not only about preserving numbers on a page; it is about safeguarding the ability of policymakers, businesses, and households to make sound decisions based on a shared understanding of economic reality.”

Roberto Rigobon, PhD ’97, is a professor of applied economics at MIT, a research associate of the National Bureau of Economic Research, a member of the Census Bureau’s Scientific Advisory Committee, and a visiting professor at IESA (Venezuela). He is co-faculty director of the MIT Sloan Sustainability Initiative and a co-founder and director of the Aggregate Confusion Project, which studies how to improve environmental, social, and governance measures.

Alberto Cavallo, MBA ’05, is a professor of business administration at Harvard Business School, a research associate at the National Bureau of Economic Research, and co-director of the Pricing Lab at Harvard’s Digital Data Design Institute. With Rigobon, Cavallo co-founded the Billion Prices Project in 2008 to expand the measurement of online inflation globally.