Trust in official economic statistics has become an increasingly salient policy issue. Across advanced economies, concerns about political pressure on statistical agencies, declining survey response rates, and the growing difficulty of measuring fast-changing economies have raised questions about the resilience of public data systems. In the US, these debates intensified after the August 2025 dismissal of Bureau of Labor Statistics (BLS) Commissioner Erika McEntarfer and accompanying public allegations that agency data had been manipulated.

Economists have long argued that uncertainty can impose substantial economic costs. Research following Bernanke (1983) and Bloom (2009) showed that heightened uncertainty reduces investment, hiring, and consumption as firms and households delay decisions. Increased uncertainty in foreign markets can have negative effects on domestic investment, consumption, and economic growth (Bali et al. 2017, Biljanovska et al. 2021). VoxEU columns have repeatedly highlighted how uncertainty shocks weaken economic activity and complicate policymaking (Bloom 2014, Baker et al. 2020, Weber et al. 2021, Salish 2026). Yet while economists have studied uncertainty extensively, much less attention has been paid to one of its underlying determinants: trust in official statistics.

Our recent research examines whether a sudden erosion in confidence in US federal statistics generated measurable economic costs. We find evidence that it did. The events surrounding the August 2025 firing of the BLS Commissioner produced a sharp increase in economic policy uncertainty, and existing estimates linking uncertainty to macroeconomic outcomes imply that the resulting loss of confidence may have reduced US GDP by roughly $20 billion.

The estimate is necessarily imprecise, but the broader implication is clear: trustworthy federal statistics are valuable economic infrastructure. This is finding is particularly relevant for democracies, which tend to maintain high-performing statistical agencies as a part of their economic infrastructure (Di Gennaro 2024).

Why statistical credibility matters

The BLS produces many of the indicators that underpin economic decision-making in the US, including the unemployment rate, monthly payroll employment estimates, wage measures, and the Consumer Price Index. These data affect monetary policy, financial markets, public spending, wage negotiations, and private-sector investment decisions.

Their importance extends beyond government. A survey of business economists by the National Association for Business Economics (Hughes-Cromwick and Coronado 2019) found that 95% viewed government statistics as important to their work, with labour market indicators ranked among the most important inputs into forecasting and planning.

Reliable statistics create value partly because they reduce uncertainty. Firms can invest more confidently when inflation and labour market conditions are measured consistently and credibly. Households can make longer-term decisions with greater confidence about wages, employment prospects, and prices. Policymakers can calibrate fiscal and monetary policy more effectively.

These benefits resemble other forms of public infrastructure: they are diffuse, economy-wide, and difficult to price directly. But one channel through which trustworthy statistics generate value — reducing uncertainty — can be measured indirectly.

Measuring the uncertainty shock

To estimate the economic consequences of the August 2025 events, we use the Economic Policy Uncertainty (EPU) Index developed by Baker et al. (2016). The index tracks the share of newspaper articles discussing economic policy uncertainty and has become a widely used measure of uncertainty shocks. It rose sharply during episodes such as the 2008 Global Financial Crisis, the 2020 COVID-19 pandemic, and periods of fiscal brinkmanship in the US in 2011 and 2013.

Following the dismissal of the BLS Commissioner on 1 August 2025, the EPU Index increased sharply (Figure 1). Comparing the week before the announcement (25–31 July) with the week after (1–7 August), the index rose by approximately 127 points, an increase of more than 50%.

Figure 1 Economic Policy Uncertainty Index around 1 August 2025 firing

Note: This figure displays the daily Economic Policy Uncertainty Index in the week before and after August 1. The solid black line is the raw daily index and the dashed lines on either side of August 1 represent the average EPU index in the week before and week after August 1 respectively.
Source: https://policyuncertainty.com/media/All_Daily_Policy_Data.csv, retrieved 27 October 2025.

Not all of that increase can be attributed solely to concerns about statistical credibility resulting from the dismissal. The same day also included major downward revisions to payroll employment estimates and the resignation announcement of Federal Reserve Governor Adriana Kugler. To isolate the portion plausibly associated with reduced trust in federal statistics, we adjust the estimate using media references specifically linked to the BLS controversy and the employment revisions.

Using this more conservative approach, we estimate that the erosion of confidence in BLS independence and data integrity increased the EPU Index by roughly 22 points, or about 9%.

Translating uncertainty into economic costs

A large literature finds that increases in uncertainty depress economic activity. Bernanke (1983) argued that uncertainty causes firms to postpone irreversible investment decisions. Bloom (2009) showed that uncertainty shocks reduce hiring and investment, while subsequent work has linked policy uncertainty to declines in output, employment, and productivity growth.

Applying estimates from this literature to the observed increase in uncertainty suggests that the decline in trust associated with the August 2025 events may have reduced GDP by roughly $20 billion.

This estimate should be interpreted carefully. It captures only one channel through which trustworthy statistics matter: the effect operating through increased uncertainty. It does not include other important benefits of federal statistics, such as improving labour market matching, informing productivity measurement, facilitating accurate inflation adjustments, or enabling evidence-based policymaking.

Even so, the comparison with agency budgets is striking. The FY2025 appropriation for the BLS was approximately $704 million. Our estimate therefore implies that the marginal erosion in trust associated with the August 2025 events may have imposed economic costs many times larger than the agency’s annual budget.

Importantly, the public did not entirely lose confidence in BLS data. Financial markets, businesses, and policymakers continued to rely heavily on official statistics. The estimated losses therefore reflect only a partial decline in trust over a short period.

Temporary disruption or lasting damage?

One important question is whether such uncertainty shocks merely delay activity or instead generate more persistent economic losses.

Some economic activity postponed during periods of uncertainty may eventually occur later. But there are reasons to believe that at least part of the effect could be lasting.

First, reputational damage to statistical agencies may persist beyond the immediate news cycle. Trust in official statistics is cumulative and institution-specific; once credibility is questioned publicly, rebuilding confidence can take considerable time.

Second, many forms of investment are not simply deferred. Decisions involving research and development, worker training, organisational restructuring, or new business formation may be cancelled altogether rather than postponed. Intangible investment, in particular, appears especially sensitive to uncertainty.

These concerns arise at a time when many statistical agencies already face operational strains. Falling survey response rates, staffing constraints, ageing technology systems, and funding pressures have complicated the production of high-quality economic statistics across advanced economies.

In the US, a recent report by the American Statistical Association (Auerbach et al. 2024) warned that federal statistical agencies remained vulnerable to political interference because of insufficient protections for professional autonomy. Meanwhile, the BLS budget has declined substantially in real terms since 2010, limiting its ability to modernise surveys and invest in new statistical methods.

Statistical agencies as economic infrastructure

Policy debates about infrastructure typically focus on roads, ports, energy systems, or broadband networks. But statistical agencies also provide foundational infrastructure for modern economies. 

In the US, the BLS, the Census Bureau, the Bureau of Economic Analysis, and related agencies generate information that allows markets and governments to function more effectively. Their outputs guide monetary policy, shape fiscal decisions, support private-sector planning, and improve public accountability.

The benefits of these institutions are difficult to observe precisely because they are embedded throughout economic decision-making. But the events of August 2025 suggest that the costs of undermining confidence in official statistics can be substantial and immediate.

Protecting the credibility, independence, and technical capacity of federal statistical agencies is therefore not merely an administrative concern. It is a consequential economic one.

References

Auerbach, J, C M Bowen and C F Citro (2024), The Nation’s Data at Risk: Meeting America’s Information Needs for the 21st Century, American Statistical Association.

Baker, S, N Bloom and S Davis (2016), “Measuring economic policy uncertainty”, Quarterly Journal of Economics 131(4): 1593–1636.

Baker, S, N Bloom, S Davis and S Terry (2020), “COVID-induced economic uncertainty”, NBER Working Paper 26983.

Bernanke, B (1983), “Irreversibility, uncertainty, and cyclical investment”, Quarterly Journal of Economics 98(1): 85–106.

Balli, F, G S Uddin, H Mudassar and S-M Yoon (2017), “Cross-country determinants of economic policy uncertainty spillovers”, Economics Letters 156: 179–183.

Biljanovska, N, F Grigoli and M Hengge (2021), “Fear thy neighbor: Spillovers from economic policy uncertainty”, Review of International Economics 29(2): 409–438.

Bloom, N (2009), “The impact of uncertainty shocks”, Econometrica 77(3): 623–685.

Bloom, N (2014), “Fluctuations in uncertainty”, VoxEU.org, 20 November. 

Bloom, N, E L Groshen, D Hobbs and M R Strain (2026), “The value of reliable statistics”, NBER Working Paper 35135.

Hughes-Cromwick, E and J Coronado (2019), “The value of US government data to US business decisions”, Journal of Economic Perspectives 33(1): 131–146.

Di Gennaro, L (2024), “Is there a quantitative relationship between democracy and official statistics?”, Statistical Journal of the IAOS 40(3): 511–519.

Salish, M (2026), “The impact of uncertainty on corporate borrowing in the euro area”, VoxEU.org, 10 January. 

Weber, M, G Kenny, D Georgarakos, Y Gorodnichenko and O Coibion (2021), “The effect of macroeconomic uncertainty on household spending”, VoxEU.org, 31 July.