Growing Reliance on CPI Imputation: A Closer Look

The US Bureau of Labor Statistics (BLS) revealed a significant surge in the proportion of prices within the August Consumer Price Index (CPI) derived from 'different-cell imputation', reaching 36%. This figure marks a jump from 32% in July and the highest level since record-keeping began in 2019. This substantial increase, nearly a 300% rise since February 2022 when the rate was just 9%, highlights a concerning degradation in the data collection system.

Concerns Over Imputation Methodology

The BLS employs 'different-cell imputation' when price data for a specific good or service is missing in a particular region, substituting it with prices from other areas. However, economists point out that this methodology can distort regional price differences, particularly in categories like housing and transportation that are highly localized. Consequently, estimation errors can be magnified.

Challenges Facing the Bureau of Labor Statistics

The BLS has faced significant challenges due to budget cuts and political interference. The Trump administration's budget reduction policies led to a substantial weakening of the BLS's data collection capabilities. The BLS staff size has decreased by at least 20% since Trump took office, and the proposed fiscal year 2026 budget suggests further cuts of 8% in funding and human resources. These constraints have led the BLS to suspend CPI sample collection in three metropolitan areas (Buffalo, Lincoln, and Provo) beginning in June 2025, in addition to reducing sampling frequency by an average of 15% in 72 other areas. While the BLS claims these adjustments have an impact of 'less than 0.01 percentage point' on the CPI, independent analyses suggest that the areas where sampling was suspended cover approximately 5% of the US consumer market, mostly smaller and mid-sized cities whose price fluctuations may differ significantly from larger cities.

Political Interference and its Impact

Persistent pressure from the Trump administration further exacerbated the BLS's difficulties. In July 2025, after the agency reported a downward revision in employment data for May and June by 258,000 jobs, Trump alleged via social media that the data was 'manipulated' and demanded a congressional investigation. Despite the lack of evidence to support this claim, the action led to low morale at the BLS, and some veteran statisticians opted for early retirement. In July 2025, Trump fired BLS Commissioner Erika McEntarfer and appointed conservative economist E.J. Antoni as her replacement, drawing sharp criticism for politicizing the agency. Nancy Mann, former president of the American Statistical Association, warns that "when political interference supplants scientific methods, the credibility of economic data collapses. Investors and businesses need accurate inflation data to make decisions, and the current BLS appears incapable of providing it."

Risks of Sample Bias

The BLS estimates CPI using two models: 'family unit' and 'different unit'. The former relies on estimating changes in prices of similar goods in the same area, while the latter relies on calling data from different areas. For example, if furniture prices are missing in a city, the system might use data from another city. Omair Sharif, president of Inflation Insights, points out that "if data collection continues to be suspended in certain areas, the proportion of 'different cell estimates' will inevitably rise. This method is theoretically possible, but in practice it may mask regional economic differences, especially in highly localized service areas such as real estate and education." According to his estimates, about 19% of CPI prices in August 2025 were estimated due to the suspension of sampling, a nearly 270% increase from 5.1% at the end of 2022. Alan Detmeister, an economist at UBS, warns that the number of CPI price quotes has been steadily declining since its peak a decade ago, which could lead to a significant increase in data volatility. "Overall, this is reducing the quality of the CPI and its ability to track true inflation," he said at a media conference call in July.

Institutional Response and Long-Term Impacts

Faced with this crisis, the Department of Labor's Office of Inspector General launched a comprehensive review of the data collection challenges at the BLS in September 2025, focusing on the impact of budget cuts, staff layoffs, and methodological adjustments on economic indicators. The BLS maintains that its statistical methods remain 'consistent with international standards' and emphasizes that the impact of suspending sampling has been 'adequately controlled' through historical data simulation. However, the market is skeptical about this. After the release of the August CPI, volatility in the dollar index increased in the short term, and federal funds rate futures showed that traders were divided over the Federal Reserve's path to cutting interest rates. Jan Hatzius, chief economist at Goldman Sachs, points out that "when 36% of price data relies on imputation, it becomes difficult for the market to believe the validity of inflation readings. This may force the Federal Reserve to rely more on other indicators, such as personal consumption expenditure and employment data, when formulating policies.