Tara Sinclair on Real-time Economic Analysis and the Fed's Upcoming Framework Review

Episode Summary

Executive Summary: Tara Sinclair discussed her Treasury experience during a period of intense macro volatility, the limits of official and private-sector real-time data, and how data revisions can materially alter policy judgments. She also weighed in on labor-market mismatch, the role of AI and big data, and the Fed’s upcoming framework review, arguing for humility, better communication, and a more explicit discussion of tools and forward-looking policy.

Main Topics: Treasury macro policy and the DAS role: Sinclair described the pace and responsibilities of serving as Deputy Assistant Secretary for Macroeconomics, emphasizing rapid-turnaround analysis, daily markets meetings, and reliance on career staff during crises like the March 2023 banking turmoil. Real-time data, revisions, and statistical fragility: The conversation stressed the trade-off between timeliness and accuracy in official data, the challenges of survey response decline and budget constraints at statistical agencies, and how revisions—especially seasonal factor changes—can reshape policy narratives. Nowcasting and private-sector data: Sinclair compared official releases with nowcasts such as the Atlanta Fed GDPNow and weekly indicators, explaining that nowcasting is useful but still constrained by seasonality and model uncertainty. She argued private data complements, but cannot replace, government statistics. Labor-market mismatch and post-pandemic reallocation: She presented research on occupation mismatch across OECD countries using Indeed data, arguing that labor markets are not homogeneous and that pandemic-era separation and rematching may have improved longer-run allocation in some cases. Fed framework review and dual mandate interpretation: Sinclair largely supported the current shortfalls framing as aligned with the dual mandate, but criticized the narrowness of the review and called for more discussion of tools, communication, and external evaluation. She emphasized that maximum employment is inherently unobservable and uncertain. AI, big data, and the future of macro analysis: She argued AI and private data can improve forecasting and analysis, but cannot eliminate the need for official statistical systems, human judgment, or policy accountability. She also floated the possibility of more stable, rules-like policy if forecasting remains unreliable.

Key Arguments: Treasury policy work is dominated by speed and crisis response, not academic-style long-horizon research; career staff provide continuity while political appointees need quick syntheses. Official macro data are essential but inherently imperfect because they trade off timeliness and accuracy and are subject to revision, especially around crises and structural changes. Statistical agencies face a worsening data environment: lower response rates, survey fatigue, smaller real budgets, and growing populations to track. CPS funding cuts are especially damaging because the survey underpins the unemployment rate, employment-to-population ratios, and demographic labor-market analysis. Real-time economics should treat preliminary releases as the relevant policy information set rather than revised data that were unavailable when decisions were made. Nowcasting is more useful than longer-horizon forecasting for understanding the current economy, though it still suffers from seasonal-adjustment and specification problems. Private-sector data and AI are complementary to government data, but they are not stable substitutes because firms change products, contracts expire, and data series can disappear. Labor markets exhibit substantial mismatch by occupation, geography, and industry; many unemployed workers are not close substitutes for any open vacancy. Pandemic-era labor reallocation may have been beneficial in the U.S. and Canada because workers separated and found better matches rather than being artificially retained in old jobs. The Fed’s shortfalls approach to employment is broadly consistent with the dual mandate, but maximum employment is not directly measurable and should not be over-specified. The Fed should consider a broader external review of tools and communication, not just internal strategy discussions, because framework choices affect policy effectiveness and accountability. AI may help improve forecasting and inference, but it will not remove uncertainty or replace the need for a stable, explainable policy framework.

Data Points: Treasury role duration: 2022 to 2024 - Sinclair served as Deputy Assistant Secretary for Macroeconomics in the U.S. Treasury during this period. DAS meeting time: 9 a.m. - Daily Treasury markets meeting where macro data and overnight market developments were discussed. Macro data release time: 8:30 a.m. - She noted that major macroeconomic indicators often arrive almost every morning at 8:30. CPS sample size: Reduced - Sinclair said the Current Population Survey has already had to cut back on sample size because of staffing and funding constraints. Framework review cadence: Every 5 years - The Fed’s consensus statement calls for a thorough public review of strategy, tools, and communication practices every five years. Fed inflation target: 2% - She stated that the Fed’s 2% inflation target is unlikely to change in the upcoming review. RTD conference start: About 2001 - Sinclair said the Real-Time Economics / Real-Time Data conference has been going on since around 2001. BEA comprehensive update frequency: Every 5 years - She referenced the Bureau of Economic Analysis’ five-year comprehensive updates to GDP and NIPA accounts. OECD countries in mismatch research: Multiple OECD countries - She described a paper using comparable Indeed data across OECD economies to study occupation mismatch. Nominal GDP discussion example: Quarterly and weekly nowcasts - The conversation contrasted standard GDP reporting with weekly economic indicators and nowcast models like GDPNow.

Pivotal Quotes: "There is no typical day for a DAS." — Tara Sinclair: She was explaining the unpredictable, high-speed nature of Treasury macro work. "The maximum level of employment is a broad-based and inclusive goal that is not directly measurable and changes over time owing largely to non-monetary factors that affect the structure and dynamics of the labor market." — David Beckworth (reading the Fed statement): This framed the discussion of how the Fed defines maximum employment in its consensus statement. "I think anything that the Fed can do that creates more jobs that does not create more inflation is something that they should actively do." — Tara Sinclair: She summarized her view of the Fed’s dual mandate and labor-market goals.

Implications: The episode suggests policy makers should treat real-time data as imperfect, invest in core statistical infrastructure, and be cautious about overclaiming what models or AI can know. For the Fed, it argues for a broader, more transparent framework review and pragmatic, accountability-focused policy.