Earnings Call, Webinar
How Reliable is Economic Data?
- Recent focus on economic data reliability stems from three primary catalysts:
- Significant revisions to the U.S. payroll survey.
- High-profile funding cuts affecting the CPI survey and other statistical agencies.
- The August dismissal of the Bureau of Labor Statistics (BLS) Commissioner by President Trump.
- Structural drivers for the perceived deterioration in data quality include:
- Pandemic aftermath: Unusual economic swings rendered standard seasonal adjustments less effective, increasing data noise.
- Chronic underfunding: Statistical agency funding has stagnated or declined in real terms over the last 10–20 years, limiting capacity to adapt to evolving economic dynamics.
- Declining survey responsiveness: Response rates have fallen long-term, accelerating post-pandemic, with participation increasingly skewed toward lower socioeconomic and older demographics in the U.S. and U.K.
- Regarding recent payroll survey revisions:
- Revisions are inherent to the statistical process, balancing the need for timeliness against accuracy.
- Initial employment reports typically include data from roughly two-thirds of employers; the remaining third reports later to reach 94–95% coverage.
- Large, unidirectional negative revisions often occur during economic turning points when struggling companies delay reporting, potentially biasing early data.
- Proposals to shift monthly data to quarterly reporting are rejected by BLS officials, as market participants and policymakers rely heavily on the timeliness of monthly indicators despite known revision risks.
- Operational impacts of current staffing and funding cuts at the BLS:
- The agency faces staff reductions of approximately 15–20%, compounded by an active hiring freeze.
- Approximately one-third of senior leadership positions are currently vacant.
- Operational resilience is compromised as remaining staff are reassigned to fill gaps, increasing the risk of errors or missed data quality controls.
- Modernization efforts are stalling due to resource reallocation, and existing price indices are losing granularity as data collection becomes more difficult.
- While errata reports have not yet increased, the capacity to correct mistakes without impacting future program scope is diminishing.
- Perspectives on political influence and data integrity:
- BLS Safeguards: Former Commissioner Erica Groshen states the data production process is automated and factory-like, with career civil servants insulated from political direction; the Commissioner sees data only after finalization and cannot alter specific values.
- Threats to Independence: Concerns persist regarding proposed changes to "Schedule P" (or similar mechanisms) that could convert career civil servants into at-will positions, potentially eroding the agency's institutional culture and independence.
- Arthur Laffer's View: Former Reagan administration economist Arthur Laffer disputes political tampering motives, attributing data issues to technical incompetence and seasonal adjustment complexities rather than conspiracy; he supports leadership shakeups as a means to reevaluate and improve processes.
- Empirical evidence on global data quality trends (Goldman Sachs Research):
- Response Rate Errors: The standard error for the U.S. JOLTS survey is estimated to be roughly 80% higher than the 2002–2013 average due to sharp contractions in firm participation.
- Inflation Measurement: The standard error for monthly CPI is projected to double following budget-driven cutbacks in data collection.
- International Trends: Australia and the U.K. have observed rising standard errors in retail sales surveys.
- Data Divergence: Significant gaps have emerged between household and payroll employment measures in the U.S. and U.K., suggesting potential sampling inaccuracies.
- Revisions Consistency: Despite quality concerns, the frequency and magnitude of data revisions have not increased notably, contradicting some narratives of systemic breakdown.
- Historical and economic implications of unreliable data:
- Former BLS advisor Alberto Cavallo cites Argentina's experience where political pressure on the statistical agency destroyed trust in official statistics, leading to prolonged capital flight and a loss of credibility that took years to recover.
- When official data is distrusted, market participants tend to assume the worst, creating asymmetric expectations and distorting inflation-related behaviors.
- Restoring trust is a slow process, and the costs to the government and economy can be substantial if data reliability continues to degrade.
- Forward-looking assessment:
- Goldman Sachs economists maintain that while global data quality has deteriorated, the current level of degradation does not yet render economic data useless for policy and investment decisions.
- However, the trajectory remains concerning if staffing cuts and political pressures are not addressed, as the risk of a total breakdown in statistical credibility increases.
- There is a cautious optimism that U.S. institutional checks and balances may prevent a slide into the chaotic environment seen in Argentina, provided the current "shaky" conditions are stabilized.