International Monetary Fund raises recession odds to 40% amid trade tensions; BLS reports 151,000 new jobs created in February 2025 as unemployment reaches 4.1%

The recent implementation of U.S. tariffs and the global trade war prompted several economists and research institutions to increase their likelihoods of a near-term recession. The International Monetary Fund, for example, recently increased the 12-month recession odds from 25 to 40 percent, citing rising tariff-related macroeconomic risk. Similarly, J.P. Morgan economists raised their recession likelihood from 40 to 60 percent for the same reason. Against the backdrop of rising recession concerns, employer hiring trends could provide early insights into the likelihood of an impending economic downturn.

Recessions are often characterized by spikes in unemployment and broad declines in economic activity. Although waves of layoffs often capture headlines during recessions, research finds that slowdowns in hiring and declines in the job-finding rate play a larger role in driving increases in unemployment. Because of deep connection between hiring and the unemployment rate, planned hiring rates revealed in survey data may provide one of the strongest early indicators of an economic downturn.

A survey by the National Federation of IndependentLabor Quality Remains a Headwind for Job Growth Based on 508 respondents to the March survey of a random sample of NFIB’s member firms, surveyed through 3/27/2025. EMBARGO 1 PM THURSDAY The U.S. Bureau of Labor Statistics (BLS) reported that in February, 151,000 jobs were created in the U.S. economy. Revisions to estimates for December and January totaled 2,000 lower than previously reported. The unemployment rate rose slightly to 4.1 percent. Employment grew in health care, financial activities, and transportation and warehousing, while federal government employment declined. Specifically looking at small businesses, NFIB’s March Small Business Economic Trends survey found that 40 percent (seasonally adjusted) of all owners reported job openings they could not fill in the current period, up 2 points from February. Thirty-three percent have openings for skilled workers (up 2 points) and 13 percent have openings for unskilled labor (unchanged). Job openings were the highest in the construction, transportation, and manufacturing sectors, and the lowest in the agriculture and wholesale sectors. Job openings in construction were up 10 points from last month (weather), and up 12 points from March 2024. Also notable, job openings in the transportation sector rose 23 points from the prior month to 53 percent. More firms reduced their employment levels than increased them (14.4 percent reduced while 9.6 percent increased). Industry- Percent with Job Openings Industry Mar. 2025 Mar. 2024 Construction 56% 44% Transportation 53% 77% Manufacturing 47% 27% Services 38% 44% Retail 35% 41% Finance 34% 13% Professional services 31% 39% Agriculture 21% 26% Wholesale 20% 16% A seasonally adjusted net 12 percent of owners plan to create new jobs in the next three months, down 3 points from February. The last time hiring plans were this low was April 2024. Job creation plans are in weak territory compared to recent history. Overall, 53 percent of owners reported hiring or trying to hire in March, unchanged from February. Forty-seven percent (87 percent of those hiring or trying to hire) of owners reported few or no qualified applicants for the positions they were trying to fill (down 1 point). Twenty-six percent of owners reported few qualified applicants for their open positions (down 1 point) and 21 percent reported none (unchanged). The percent of small business owners reporting labor quality as their top small business operating problem was unchanged from February at 19 percent. Labor costs reported as the single most important problem for business owners fell 1 point from February to 11 percent, only 2 points below the highest reading of 13 percent reached in December 2021. Seasonally adjusted, a net 38 percent reported raising compensation, up 5 points from February. Small business owners are feeling pressured to retain and attract employees. A net 19 percent (seasonally adjusted) plan to raise compensation in the next three months, up 1 point from February. Overall, wage cost increases continue to pressure the bottom line for owners. Main Street job openings hold firm as qualified workers continue to be in short supply. However, job creation plans have started to fade over the last few months. Public sector (government, social services, education, and healthcare) job growth remained solid at pre-DOGE levels, but data releases in the next few months should start reflecting recent layoffs. Finding qualified workers though remains the major headwind for stronger job growth on Main Street. The new administration is aggressively addressing the rapid growth of government employment and spending, muddying the employment picture (many workers released will be paid through September). Overall, economic growth appears to be slowing, but at a slow pace. While large firms are shrinking their employment, small businesses are still trying to fill job openings. Wage increases remain strong, putting pressure on prices. The bottom line for all of this will become clearer over the next few months. Business dating back to 1971 measures the share of small businesses planning to create jobs in the next three months. The benefit of this survey is that it captures several recessions, highlighting a systemic relationship between planned hiring and a subsequent recession. The data shows steep declines in planned hiring during recessions and several cases of planned hiring dropping in the months preceding a recession. The most recent data shows two consecutive months of decline—a concerning sign, considering the relative stability of the series.

Other surveys have confirmed the recent decline in expected hiring. The Federal Reserve Bank of Dallas collects some of the most detailed data on firms’ projected hiring activity. Although the data are collected only for Texas firms, surveys from other regions of the U.S. show consistency in recent trends. The data show that after the wave of post-pandemic stimulus peaked in the latter half of 2021, planned hiring declined steadily through the end of 2024. By the start of 2025, the outlook appeared to rebound before falling sharply again as tariff concerns escalated.

A national-level survey published by the Federal Reserve Bank of AtlantaSurvey of Business Uncertainty Monthly Report April 2025 Brent Meyer, Jose Maria Barrero, Nicholas Bloom, Steven J. Davis, Kevin Foster, and Emil Mihaylov SBU Based on survey responses from 14-25 April 2025 Headline Results April 2025 Survey of Business Uncertainty 1. Sales revenue growth expectations have declined in recent months. (Slide 4) 2. Firms remain more uncertain about future sales growth than before the pandemic. (Slide 4) 3. Forty-five percent of firms anticipate scaling back on investment plans and 40 percent anticipate scaling back on hiring due to policy uncertainty. (Slides 7) About the Survey The Survey of Business Uncertainty (SBU) is fielded each month by the Federal Reserve Bank of Atlanta. The SBU questionnaire goes to about 1500 panel members, who occupy senior finance and managerial positions at U.S. firms. We contact panel members each month by email, and they respond via a web-based instrument. Survey questions pertain to current, past, and future outcomes at the respondent’s firm. Our primary objective is to elicit the respondent’s subjective forecast distributions over own-firm future sales growth rates and employment levels. We also ask special questions on timely topics. For more information on survey design and methodology, please refer to the resources on the SBU page and “Surveying Business Uncertainty,” published in the Journal of Econometrics and also available as NBER Working Paper 25956. Nominal sales growth has slowed considerably over the past two years. Recent employment growth is in line with pre-pandemic growth. January 2017–April 2025 NOTE: Calculated using monthly data through April 2025. Realized growth rate series for sales revenue and employment are activity-weighted averages of firms’ reported (look-back) growth rates over the past year (specifically, the previous four quarters for sales revenue and previous 12 months for employment). NOTE: The chart shows smoothed series. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. For more information, see “Surveying Business Uncertainty” by David Altig, Jose Maria Barrero, Nick Bloom, Steven J. Davis, Brent Meyer, and Nick Parker, NBER Working Paper No. 25956, February 2020. The vertical dashed line shown in the plot marks the start of the COVID-19 pandemic. 3 Sales revenue growth expectations have declined in recent months. However, firms remain more uncertain about future revenue growth than they were before the pandemic. January 2017–April 2025 NOTE: The charts show smoothed series. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. For more information, see “Surveying Business Uncertainty” by David Altig, Jose Maria Barrero, Nick Bloom, Steven J. Davis, Brent Meyer, and Nick Parker, NBER Working Paper No. 25956, February 2020. The vertical dashed lines shown in the plots mark the start of the COVID-19 pandemic. 4 Expected employment growth has decreased in recent months. Uncertainty about employment growth has returned to pre-pandemic levels. January 2017–April 2025 NOTE: The charts show smoothed series. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. For more information, see “Surveying Business Uncertainty” by David Altig, Jose Maria Barrero, Nick Bloom, Steven J. Davis, Brent Meyer, and Nick Parker, NBER Working Paper No. 25956, February 2020. The vertical dashed lines shown in the plots mark the start of the COVID-19 pandemic. 5 The distribution of sales growth rates across firms remains wider than before the pandemic. January 2017–April 2025 NOTES: Calculated using monthly data through April 2025. The chart shows smoothed series. Lines show percentiles of the activity-weighted distribution of firm-level sales growth rates over the past year. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. 6 Forty-five percent of firms reported that they are scaling back investment and 40 percent of firms are scaling back hiring due to policy uncertainty. Question: Has uncertainty about tariffs, taxes, government spending, monetary policy, or regulation affected your firm’s plans for [hiring, investment] over the next 6 months? Note: These questions were fielded in the April 2025 SBU survey wave. 7 SBU panelists adjusted downward their long-run expectations about the federal funds rate after the March 2025 FOMC meeting Question: We would like to ask you about the interest rate set by the Federal Reserve's Federal Open Market Committee (FOMC). This interest rate is often referred to as the Federal Funds Rate. What do you think is the current Federal Funds Rate? Additionally, what do you think the Federal Funds Rate will be, three months from now, one year from now, and three years from now? Note: Results are equally weighted. These questions were fielded in the March 2025 SBU survey wave. 8 Foreign sales as percent of total revenues from firms' US operations Question: Now think about foreign sales and input purchases for your firm. What percent of revenues from your firm's U.S. operations are due to foreign sales? Note: These questions were fielded in the March 2025 SBU survey wave. 9 Share of foreign-sourced non-labor inputs used in firms' U.S. operations Question: What percent of non-labor inputs used by your firm's U.S. operations are sourced from abroad? Note: These questions were fielded in the March 2025 SBU survey wave. 10 Percent of firms' production occurring outside of the U.S. Question: Now think about foreign production, sales, and input purchases for your firm. What percent of your firm's production occurs outside the United States? Note: These questions were fielded in the March 2025 SBU survey wave. 11 Percent of firms' global revenues due to foreign sales Question: What percent of your firm's global revenues are due to foreign sales? Note: These questions were fielded in the March 2025 SBU survey wave. 12 Foreign sales as percent of total revenues from firms' US operations Question: What percent of revenues from your firm's U.S. operations are due to foreign sales? Note: These questions were fielded in the March 2025 SBU survey wave. 13 Share of foreign-sourced non-labor input supplies used in firms' U.S. operations Question: What percent of non-labor input supplies for your firm's U.S. operations does your firm source from abroad? Note: These questions were fielded in the March 2025 SBU survey wave. 14 Appendix: Technical Information Computing Moments of the Firm-Level Subjective Forecast Distributions Subjective Expectations and Uncertainty Indices Topic-specific Expected Excess Reallocation Indices We calculate first and second moments of the subjective growth rate distributions of We construct a monthly activity-weighted expectations (first-moment) index for We construct forward-looking indices of excess job and sales revenue reallocation. employment and sales revenue over the next 12 months or four quarters, as employment growth and sales growth looking one year ahead. We also construct a These series measure the volume of cross-firm reallocation in economic activity above appropriate. Following standard practice in the literature on business-level dynamics, monthly activity-weighted uncertainty (second-moment) index for the employment the reallocation required to support aggregate growth. For ease of exposition, we often we calculate the growth rate of x from t–1 to t as 𝑔𝑡 = 2(𝑥𝑡 − 𝑥𝑡−1)/ 𝑥𝑡 + 𝑥𝑡−1 .* growth and sales growth looking one year ahead. refer to these as simply “reallocation rates”: Employment • First, in each month t, we compute the activity-weighted average of own-firm 𝐶𝐸𝑚𝑝 = firm’s current employment level, as reported by the respondent • In month t, the index for employment (sales) takes a value equal to the activity- expected gross job creation and destruction rates, which boils down to the weighted average of subjective mean employment (sales) growth rates looking 𝐹𝐸𝑚𝑝𝑖 = employment 12 months hence in scenario 𝑖, for 𝑖 = 1, 2, 3, 4, 5 activity-weighted average of the absolute value of subjective mean growth one year hence ( Mean(𝐺𝑟) ), averaging across all firms responding that month. 𝑝𝑖 = rates |Mean(𝐸𝐺𝑟) |. the associated probabilities, 𝑖 = 1, 2, 3, 4, 5 We compute these subjective mean growth rates as described on slide 3, and winsorize them at the first and 99th percentiles before using them to construct the index. • Then , in each month t, we compute the absolute value of the activity weighted Scenario-Specific Growth Rates average of own-firm expected employment growth Mean(𝐸𝐺𝑟) . This is 𝐸𝐺𝑟𝑖 = 2(𝐹𝐸𝑚𝑝𝑖−𝐶𝐸𝑚𝑝)/(𝐹𝐸𝑚𝑝𝑖+𝐶𝐸𝑚𝑝), 𝑖 = 1, 2, 3, 4, 5 effectively the absolute value of the employment growth expectations index in • The month-t index of year-ahead subjective uncertainty for employment (sales) month t. growth is the activity-weighted mean of (SD (𝐺𝑟) ) values across firms First and Second Moments of the Subjective Growth Rate Forecast Distribution responding in month t. We compute these subjective standard deviations over growth rates as described on slide 3, and winsorize them at the first and 99th • We then obtain the expected job reallocation rate index value for month t by Mean(𝐸𝐺𝑟) = σ5 𝑖=1 𝑝𝑖 𝐸𝐺𝑟𝑖 percentiles before inputting them into the index construction formula. subtracting the outcome of the second bullet from the first. Letting 𝑤𝑖𝑡 be firm Var(𝐸𝐺𝑟) = σ5 𝑖=1 𝑝𝑖 𝐸𝑚𝑝𝐺𝑟𝑖 − Mean(𝐸𝐺𝑟) 2 𝑖’s activity weight in month 𝑡, SD(𝐸𝐺𝑟) = Var(𝐸𝐺𝑟) • When constructing first- and second-moment employment growth indexes, we weight firm i’s subjective mean growth rate expectation and uncertainty by the Sales Revenue average of its month-t employment (CEmpit) and its expected employment level 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝐽𝑜𝑏 𝑅𝑒𝑎𝑙𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒𝑡 = ෍ 𝑤𝑡 ⋅ |Mean(𝐸𝐺𝑟) | − ෍ 𝑤𝑡 ⋅ Mean(𝐸𝐺𝑟) 𝐶𝑆𝑎𝑙𝑒 = firm’s sales revenue in the current quarter, as reported by the respondent (𝐸𝐸𝑚𝑝𝑖𝑡). We top-code these weights at 500 to diminish the influence of outliers 𝑖 𝑖 among very large firms. 𝐹𝑆𝑎𝑙𝑒𝐺𝑟𝑖 = respondent’s scenario–specific sales growth rate from now to four quarters • Analogously, the expected sales revenue reallocation rate index in month t is hence, 𝑖 = 1, 2, 3, 4, 5 the difference between the activity-weighted average of absolute expected • When constructing first- and second-moment sales revenue growth indexes, we sales growth rates, minus the absolute value of the average activity-weighted 𝑝𝑖 = the associated probabilities, 𝑖 = 1, 2, 3, 4, 5 weight firms i’s subjective mean growth rate expectation and uncertainty by the growth rate: Implied Future Sales Level average of its month-t sales revenue (CSaleit) and its expected sales level 𝐹𝑆𝑎𝑙𝑒𝐺𝑟 (ESaleit). We winsorize these activity-weights at the 1st and 80th percentile. 𝐹𝑆𝑎𝑙𝑒𝑖 = 1 + 𝑖 𝐶𝑆𝑎𝑙𝑒, 𝑖 = 1, 2, 3, 4, 5 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑅𝑒𝑎𝑙𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒 𝐹𝑜𝑟 𝑆𝑎𝑙𝑒𝑠 𝑅𝑒𝑣𝑒𝑛𝑢𝑒𝑡 100 • Finally, we smooth our topic-specific indices by taking a moving average. We set = ෍ 𝑤𝑡 ⋅ |Mean(𝑆𝑎𝑙𝑒𝐺𝑟) | − ෍ 𝑤𝑡 ⋅ Mean(𝑆𝑎𝑙𝑒𝐺𝑟) Scenario–Specific Growth Rates (re–expressing respondent growth rates to our the window for the moving average to 2 or 3 months, to match the panel structure 𝑖 𝑖 growth rate measure) of our survey. • We compute the subjective mean growth rates Mean(𝐸𝐺𝑟) and 𝑆𝑎𝑙𝑒𝐺𝑟𝑖 = 2(𝐹𝑆𝑎𝑙𝑒𝑖−𝐶𝑆𝑎𝑙𝑒𝑠)/(𝐹𝑆𝑎𝑙𝑒𝑖+𝐶𝑆𝑎𝑙𝑒) = 2𝐹𝑆𝑎𝑙𝑒𝐺𝑟𝑖/(𝐹𝑆𝑎𝑙𝑒𝐺𝑟𝑖 + 2), 𝑖 = Mean(𝑆𝑎𝑙𝑒𝐺𝑟) as described on slides 18-21, and winsorize them at the 1st 1, 2, 3, 4, 5 and 99th percentiles before using them to construct the index. First and Second Moments of the Subjective Growth Rate Forecast Distribution • Firm i’s activity weight 𝑤𝑖𝑡 is the average of its month–t employment or sales Mean(𝑆𝑎𝑙𝑒𝐺𝑟) = σ5 𝑖=1 𝑝𝑖 𝑆𝑎𝑙𝑒𝐺𝑟 level (Cempit or CSaleit) and its expected employment or sales level twelve 𝑖 months hence (𝐹𝐸𝑚𝑝𝑖𝑡 or FSaleit). We top–code these weights at 500 for Var(𝑆𝑎𝑙𝑒𝐺𝑟) = σ5 𝑖=1 𝑝𝑖 𝑆𝑎𝑙𝑒𝐺𝑟𝑖 − Mean(𝑆𝑎𝑙𝑒𝐺𝑟) 2 𝑖 employment and at the 80th percentile for sales to diminish the influence of SD(𝑆𝑎𝑙𝑒𝐺𝑟) = Var(𝑆𝑎𝑙𝑒𝐺𝑟) outliers among very large firms. 15 Appendix: Subjective Forecast Distribution of Future Sales Growth Rates at a One-Year Horizon January 2017–April 2025 NOTES: Calculated using monthly data through April 2025. The charts show smoothed series. This is a plot of the subjective distribution for the representative firm’s future sales growth rates over a 4- quarter look-ahead horizon. To calculate this distribution, we pool over all firm-level subjective forecast distributions in the indicated month and weight each firm by its activity level. Then we use the probabilities assigned to each possible future sales growth rate to obtain activity-weighted quantiles of the future sales growth rate distribution. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. 16 Appendix: Histogram of survey response frequency for the April 2025 survey wave April 2025 Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. 17 also tracks projected hiring activity. The survey measures expected employment growth rates from a large sample of U.S. firms. Recent data show strong expected employment growthSurvey of Business Uncertainty Monthly Report March 2025 Brent Meyer, Jose Maria Barrero, Nicholas Bloom, Steven J. Davis, Kevin Foster, and Emil Mihaylov SBU Based on survey responses from 10-21 March 2025 Headline Results March 2025 Survey of Business Uncertainty 1. Sales revenue growth expectations have ticked up in recent months. (Slide 4) 2. But firms remain more uncertain about future sales growth than before the pandemic. (Slide 4) 3. SBU firms don’t foresee drastic changes in working arrangements over the year ahead. (Slides 7-9) About the Survey The Survey of Business Uncertainty (SBU) is fielded each month by the Federal Reserve Bank of Atlanta. The SBU questionnaire goes to about 1500 panel members, who occupy senior finance and managerial positions at U.S. firms. We contact panel members each month by email, and they respond via a web-based instrument. Survey questions pertain to current, past, and future outcomes at the respondent’s firm. Our primary objective is to elicit the respondent’s subjective forecast distributions over own-firm future sales growth rates and employment levels. We also ask special questions on timely topics. For more information on survey design and methodology, please refer to the resources on the SBU page and “Surveying Business Uncertainty,” published in the Journal of Econometrics and also available as NBER Working Paper 25956. Nominal sales growth has slowed considerably over the past two years. Recent employment growth is in line with pre-pandemic growth. January 2017–March 2025 NOTE: Calculated using monthly data through March 2025. Realized growth rate series for sales revenue and employment are activity-weighted averages of firms’ reported (look-back) growth rates over the past year (specifically, the previous four quarters for sales revenue and previous 12 months for employment). NOTE: The chart shows smoothed series. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. For more information, see “Surveying Business Uncertainty” by David Altig, Jose Maria Barrero, Nick Bloom, Steven J. Davis, Brent Meyer, and Nick Parker, NBER Working Paper No. 25956, February 2020. The vertical dashed line shown in the plot marks the start of the COVID-19 pandemic. 3 Sales revenue growth expectations have ticked up in recent months. However, firms remain more uncertain about future revenue growth than they were before the pandemic. January 2017–March 2025 NOTE: The charts show smoothed series. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. For more information, see “Surveying Business Uncertainty” by David Altig, Jose Maria Barrero, Nick Bloom, Steven J. Davis, Brent Meyer, and Nick Parker, NBER Working Paper No. 25956, February 2020. The vertical dashed lines shown in the plots mark the start of the COVID-19 pandemic. 4 Expected employment growth has improved in recent months. Uncertainty about employment growth has returned to pre-pandemic levels. January 2017–March 2025 NOTE: The charts show smoothed series. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. For more information, see “Surveying Business Uncertainty” by David Altig, Jose Maria Barrero, Nick Bloom, Steven J. Davis, Brent Meyer, and Nick Parker, NBER Working Paper No. 25956, February 2020. The vertical dashed lines shown in the plots mark the start of the COVID-19 pandemic. 5 The distribution of sales growth rates across firms remains wider than before the pandemic. January 2017–March 2025 NOTES: Calculated using monthly data through March 2025. The chart shows smoothed series. Lines show percentiles of the activity-weighted distribution of firm-level sales growth rates over the past year. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. 6 SBU firms don’t foresee drastic changes in working arrangements over the year ahead Question: What do you expect would be the share of your firm's full-time employees in each category under the return-to-office mandate? Answers should sum to 100 Your firm's current shares are in parentheses • Paid working days at home as a percent of all working days currently: 21.2% • Paid working days at home as a percent of all working days after under the return-to-office mandate: 20.8% Paid working days at home as a percent of all working days is calculated by converting the number of days at home to a fraction of the 5-day workweek (0.3 for 1-2 days, 0.7 for 3-4 days, and 1 for 5 days) Note: Results are weighted by firm employment. These questions were fielded in the February 2025 SBU survey wave. 7 Changes in firms’ WFH stances are evenly distributed and centered on zero. Question: Currently, what share of your firm's full-time employees are in each category? Answers should sum to 100 Note: The mean and standard error are weighted by firm employment. These questions were fielded in the November 2024 and February 2025 SBU survey waves. Paid working days at home as a percent of all working days is calculated by converting the number of days at home to a fraction of the workweek (0.3 for 1-2 days, 0.7 for 3-4 days, and 1 for 5 days) 8 A significant downturn in economic conditions motivates a slight decline in WFH Question: You said that your firm would increase/reduce the share of hybrid or fully-remote employees if the unemployment rate were to double. What do you expect would be the share of your firm's full-time employees in each category under the new policy? Answers should sum to 100 Your firm's current shares are in parentheses Note: Results are weighted by firm employment. These questions were fielded in the February 2025 SBU survey wave. 9 Appendix: Technical Information Computing Moments of the Firm-Level Subjective Forecast Distributions Subjective Expectations and Uncertainty Indices Topic-specific Expected Excess Reallocation Indices We calculate first and second moments of the subjective growth rate distributions of We construct a monthly activity-weighted expectations (first-moment) index for We construct forward-looking indices of excess job and sales revenue reallocation. employment and sales revenue over the next 12 months or four quarters, as employment growth and sales growth looking one year ahead. We also construct a These series measure the volume of cross-firm reallocation in economic activity above appropriate. Following standard practice in the literature on business-level dynamics, monthly activity-weighted uncertainty (second-moment) index for the employment the reallocation required to support aggregate growth. For ease of exposition, we often we calculate the growth rate of x from t–1 to t as 𝑔𝑡 = 2(𝑥𝑡 − 𝑥𝑡−1)/ 𝑥𝑡 + 𝑥𝑡−1 .* growth and sales growth looking one year ahead. refer to these as simply “reallocation rates”: Employment • First, in each month t, we compute the activity-weighted average of own-firm 𝐶𝐸𝑚𝑝 = firm’s current employment level, as reported by the respondent • In month t, the index for employment (sales) takes a value equal to the activity- expected gross job creation and destruction rates, which boils down to the weighted average of subjective mean employment (sales) growth rates looking 𝐹𝐸𝑚𝑝𝑖 = employment 12 months hence in scenario 𝑖, for 𝑖 = 1, 2, 3, 4, 5 activity-weighted average of the absolute value of subjective mean growth one year hence ( Mean(𝐺𝑟) ), averaging across all firms responding that month. 𝑝𝑖 = rates |Mean(𝐸𝐺𝑟) |. the associated probabilities, 𝑖 = 1, 2, 3, 4, 5 We compute these subjective mean growth rates as described on slide 3, and winsorize them at the first and 99th percentiles before using them to construct the index. • Then , in each month t, we compute the absolute value of the activity weighted Scenario-Specific Growth Rates average of own-firm expected employment growth Mean(𝐸𝐺𝑟) . This is 𝐸𝐺𝑟𝑖 = 2(𝐹𝐸𝑚𝑝𝑖−𝐶𝐸𝑚𝑝)/(𝐹𝐸𝑚𝑝𝑖+𝐶𝐸𝑚𝑝), 𝑖 = 1, 2, 3, 4, 5 effectively the absolute value of the employment growth expectations index in • The month-t index of year-ahead subjective uncertainty for employment (sales) month t. growth is the activity-weighted mean of (SD (𝐺𝑟) ) values across firms First and Second Moments of the Subjective Growth Rate Forecast Distribution responding in month t. We compute these subjective standard deviations over growth rates as described on slide 3, and winsorize them at the first and 99th • We then obtain the expected job reallocation rate index value for month t by Mean(𝐸𝐺𝑟) = σ5 𝑖=1 𝑝𝑖 𝐸𝐺𝑟𝑖 percentiles before inputting them into the index construction formula. subtracting the outcome of the second bullet from the first. Letting 𝑤𝑖𝑡 be firm Var(𝐸𝐺𝑟) = σ5 𝑖=1 𝑝𝑖 𝐸𝑚𝑝𝐺𝑟𝑖 − Mean(𝐸𝐺𝑟) 2 𝑖’s activity weight in month 𝑡, SD(𝐸𝐺𝑟) = Var(𝐸𝐺𝑟) • When constructing first- and second-moment employment growth indexes, we weight firm i’s subjective mean growth rate expectation and uncertainty by the Sales Revenue average of its month-t employment (CEmpit) and its expected employment level 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝐽𝑜𝑏 𝑅𝑒𝑎𝑙𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒𝑡 = ෍ 𝑤𝑡 ⋅ |Mean(𝐸𝐺𝑟) | − ෍ 𝑤𝑡 ⋅ Mean(𝐸𝐺𝑟) 𝐶𝑆𝑎𝑙𝑒 = firm’s sales revenue in the current quarter, as reported by the respondent (𝐸𝐸𝑚𝑝𝑖𝑡). We top-code these weights at 500 to diminish the influence of outliers 𝑖 𝑖 among very large firms. 𝐹𝑆𝑎𝑙𝑒𝐺𝑟𝑖 = respondent’s scenario–specific sales growth rate from now to four quarters • Analogously, the expected sales revenue reallocation rate index in month t is hence, 𝑖 = 1, 2, 3, 4, 5 the difference between the activity-weighted average of absolute expected • When constructing first- and second-moment sales revenue growth indexes, we sales growth rates, minus the absolute value of the average activity-weighted 𝑝𝑖 = the associated probabilities, 𝑖 = 1, 2, 3, 4, 5 weight firms i’s subjective mean growth rate expectation and uncertainty by the growth rate: Implied Future Sales Level average of its month-t sales revenue (CSaleit) and its expected sales level 𝐹𝑆𝑎𝑙𝑒𝐺𝑟 (ESaleit). We winsorize these activity-weights at the 1st and 80th percentile. 𝐹𝑆𝑎𝑙𝑒𝑖 = 1 + 𝑖 𝐶𝑆𝑎𝑙𝑒, 𝑖 = 1, 2, 3, 4, 5 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑅𝑒𝑎𝑙𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒 𝐹𝑜𝑟 𝑆𝑎𝑙𝑒𝑠 𝑅𝑒𝑣𝑒𝑛𝑢𝑒𝑡 100 • Finally, we smooth our topic-specific indices by taking a moving average. We set = ෍ 𝑤𝑡 ⋅ |Mean(𝑆𝑎𝑙𝑒𝐺𝑟) | − ෍ 𝑤𝑡 ⋅ Mean(𝑆𝑎𝑙𝑒𝐺𝑟) Scenario–Specific Growth Rates (re–expressing respondent growth rates to our the window for the moving average to 2 or 3 months, to match the panel structure 𝑖 𝑖 growth rate measure) of our survey. • We compute the subjective mean growth rates Mean(𝐸𝐺𝑟) and 𝑆𝑎𝑙𝑒𝐺𝑟𝑖 = 2(𝐹𝑆𝑎𝑙𝑒𝑖−𝐶𝑆𝑎𝑙𝑒𝑠)/(𝐹𝑆𝑎𝑙𝑒𝑖+𝐶𝑆𝑎𝑙𝑒) = 2𝐹𝑆𝑎𝑙𝑒𝐺𝑟𝑖/(𝐹𝑆𝑎𝑙𝑒𝐺𝑟𝑖 + 2), 𝑖 = Mean(𝑆𝑎𝑙𝑒𝐺𝑟) as described on slides 18-21, and winsorize them at the 1st 1, 2, 3, 4, 5 and 99th percentiles before using them to construct the index. First and Second Moments of the Subjective Growth Rate Forecast Distribution • Firm i’s activity weight 𝑤𝑖𝑡 is the average of its month–t employment or sales Mean(𝑆𝑎𝑙𝑒𝐺𝑟) = σ5 𝑖=1 𝑝𝑖 𝑆𝑎𝑙𝑒𝐺𝑟 level (Cempit or CSaleit) and its expected employment or sales level twelve 𝑖 months hence (𝐹𝐸𝑚𝑝𝑖𝑡 or FSaleit). We top–code these weights at 500 for Var(𝑆𝑎𝑙𝑒𝐺𝑟) = σ5 𝑖=1 𝑝𝑖 𝑆𝑎𝑙𝑒𝐺𝑟𝑖 − Mean(𝑆𝑎𝑙𝑒𝐺𝑟) 2 𝑖 employment and at the 80th percentile for sales to diminish the influence of SD(𝑆𝑎𝑙𝑒𝐺𝑟) = Var(𝑆𝑎𝑙𝑒𝐺𝑟) outliers among very large firms. 10 Appendix: Subjective Forecast Distribution of Future Sales Growth Rates at a One-Year Horizon January 2017–March 2025 NOTES: Calculated using monthly data through March 2025. The charts show smoothed series. This is a plot of the subjective distribution for the representative firm’s future sales growth rates over a 4- quarter look-ahead horizon. To calculate this distribution, we pool over all firm-level subjective forecast distributions in the indicated month and weight each firm by its activity level. Then we use the probabilities assigned to each possible future sales growth rate to obtain activity-weighted quantiles of the future sales growth rate distribution. Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. 11 Appendix: Histogram of survey response frequency for the March 2025 survey wave March 2025 Source: Survey of Business Uncertainty conducted by the Federal Reserve Bank of Atlanta. 12 at the start of 2025, followed by a sharp decline, likely reflecting a weakening macroeconomic outlook.

The influential role of changes in hiring and job-finding to spikes in the unemployment rate during recessions underscores the importance of monitoring hiring activity. Moreover, planned hiring trends can provide an early indication of an economic downturn. Recent declines in planned hiring activity across various surveys indicate an increase in recession risk, validating the concerns expressed by several economists.