Federal Partners Panel: An Update from Agency Leaders


This is a series of keynote summaries from the virtual 60th Annual C2ER Conference and LMI Institute Forum held June 1 – 5, 2020.

New data on job openings and labor turnover at the metro level could result if BLS gets its proposed funding increase, said Commissioner William Beach of the Bureau of Labor Statistics at the first virtual C2ER/LMI Institute Annual Conference on June 5. Speaking on a panel with Dr. Mary Bohman, Acting Director of the Bureau of Economic Analysis (BEA), and Dr. Ron Jarmin, Deputy Director of the U.S. Census Bureau, Dr. Beach noted that BLS has released some experimental data from JOLTS at the state level and hopes to expand its sample to provide more granular data. The three agency directors discussed: 1) their current budget situation, 2) COVID-19 impacts their agencies, 3) data quality concerns, and 4) agency product and process innovations with Dr. Michael Horrigan, CEO of the Upjohn Institute, who moderated the discussion.

Current Budget Situation in Federal Economic Statistical Agencies

All three statistical agencies expressed satisfaction with their proposed 2021 budgets and an appreciation for modest funding increases in 2020. Dr. Jarmin explained that the Census Bureau secured the budget for this fiscal year and flexibility to use the supplemental funding provided for the decennial Census, but the agency had to re-plan the 2020 Census field operations due to the pandemic. That has pushed back all field activities and delayed decennial data collection operations continue much later into the fall than initially planned. Contingency funds available will help for data collection and enumeration, but there are proposals in Congress to provide supplemental funding to address the delays and the need to equip field staff with personal protective equipment.

Dr. William Beach noted that BLS received a $655 million appropriation and was not included in the CARES Act or other COVID-related stimulus bills. That appropriation includes partial funding for a planned move to Suitland, Maryland, in 2022 where BLS will be co-located with Census and BEA.  Still, with the proposed increase in FY2021 appropriations ($658.3 million for BLS), the agency plans to fund the remainder of the expected move costs and supplement its data collection efforts. In the President’s budget request, $7.1 million is associated with expanding the Job Openings & Labor Turnover Survey (JOLTS) by increasing the sample size to produce sample-based estimates for every state with more detailed questions. Dr. Beach expects BLS to produce a JOLTS survey and employment situation survey in the same month as the reference period, which eventually will produce near real-time supply-side and demand-side information about the dynamics of the labor market.

Dr. Beach announced new JOLTS production for the 18 Metropolitan Statistical Areas (MSAs) – which covers about 40% of all U.S. employees – in Summer 2020. By more than doubling the sample size, BLS is likely to expand the estimates to include the top 50 MSAs as well with the agency’s sample-based approach.

Dr. Mary Bohman, Acting Director, highlighted two new activities at the BEA from the FY2020: 1) production of GDP statistics for Puerto Rico, and 2) establishment of Federal Data Service (in cooperation with Census) to make more data available to researchers. She expects that BEA can publish GDP at the national/state and industry levels simultaneously from the third quarter of 2020, vastly improving the timeliness of state GDP data releases. BEA also plans to continue publishing county-level GDP, including at least 20 industry-level details. The budget for FY2021 continues to support GDP, data service, and evidence-based policymaking.

How did COVID-19 Impact Federal Statistical Agencies?

The COVID-19 pandemic impacted the federal statistical agencies dramatically in terms of field operations and data collection activities. All three agencies have moved to a 100% telework environment during the pandemic while adjusting their means of data collection. Dr. Beach explained how BLS moved its price index-side of data collection from in-store collection to non-human contact methods, including video links and emails. Although the response rate is lower, BLS did not suppress any details. The household survey for the Current Population Survey is under similar conditions. BLS stopped all personal contacts, and the response rate was about mid-60%. He also explained how BLS initiated a weekly conversation with the labor market information (LMI) shops in key states to stay abreast of data collection and quality challenges.

For the Census Bureau, moving to 100% telework specifically took a toll on the 2020 Census, closing about 250 area Census offices. Dr. Jarmin noted that the Census Bureau stopped household visits and pushed non-response follow-ups to August, a significant delay that will impact the timing of delivering the final Decennial 2020 product. While relying on self-response, the Bureau looked for a different way to reach out to respondents using social media or emails. Although the response rate was a bit down for business & household surveys, the Census Bureau has been able to push data out to users by enhancing existing statistics. Business Formation Statistics, for example, used to be a quarterly product, but the Census started to release it weekly using administrative (business application) data. Another example is the launch of the Small Business Pulse Survey and Household Pulse Survey, both measuring the impact of the pandemic on the ground.

Dr. Bohman shed light on the daily operation and cooperation with other organizations during the pandemic. 2020 is a benchmark year for GDP data, which comes from a BEA-run survey for multinational companies. As a huge consumer of the federal government’s data product, BEA spends a considerable amount of time working with other organizations and sharing data adjustments and methods. Understanding the change in the source data and documenting the impact of COVID-19 on their statistical product became the main tasks. This was made possible via weekly calls among economic agencies, including Federal Reserves, BLS, Census, and OMB Interagency Council on Statistical Policy and in-depth discussion for specific topics (e.g., Joint BEA-BLS product account).

How does Your Organization Maintain Data Quality?  

As the nation experiences remarkable changes in economic activity, concerns for data quality has risen. Dr. Bohman explained how they tried to overcome the challenges that BEA faces, emphasizing three points. First, transparency is a big part of the story. Publishing comprehensive technical notes, documentation, and FAQs for any changes helps BEA’s data transparency. For example, BEA identified and published how the federal pandemic response program impacted the National Income and Product Account. Second, good source data is key. With ongoing discussion with BLS and Census, BEA incorporated other federal statistical agencies’ thinking into their estimates. Third, long-term investment is essential. BEA also has capitalized on big data from the private sector (e.g., monthly credit card data published from private vendors). With this investment, the agency published a working paper for consumer spending identifying the consumption patterns during the pandemic.

Dr. Beach brought up measurement issues at the outset, taking the period of mid-March to the end of March as an example. Since BLS does not know precisely how many workers dropped out of the system during those two weeks, it affects the calculation of labor productivity estimates. Labor productivity is calculated by dividing the total amount of output, which BEA publishes, with the number of hours that worked. However, BLS took an innovative approach using “continuing unemployment insurance (UI) claims datasets” – maintained by Employment and Training Administration at the Department of Labor – and calibrated the change in hours worked for a different industry.

Dr. Jarmin emphasized the importance of source data improvement as well as core measurement capability. While agreeing with Dr. Bohman on the importance of long-term investment and alternative data sources, he also pointed out that federal statistical agencies need to stay apolitical. However, the private sector indeed possesses vast data that is useful, such as JP Morgan & Chase’s small business data dashboard. Federal statistical agencies’ responsibility is to triangulate different data sources and enhance their core measurement capability.

Any other innovation at your organization? 

BLS is commissioning two subject-matter expert panels on 1) retail trade and 2) Consumer Price Index (CPI). The price index last received such a review in 2002. BLS is also trying to modernize the Consumer Expenditure Survey, which provides sampling frames for the CPI, with the elimination of the paper diary form and moving toward online or app-based data collection to build more electronic data. Additionally, BLS is about to roll out a contingent workforce survey every two years, as opposed to every ten years now, after receiving a report from the Committee on National Statistics (CN Stat) at the National Academies of Sciences Engineering Medicine.

BEA has several new data products, including publication of county-level GDP, prototype outdoor recreation statistics at state-level, statistics measuring the ocean economy in partnership with National Oceanic and Atmospheric Administration, and prototype statistics on personal income distribution. While the Census Bureau provides detailed demographics around income distribution, BEA’s new product reveals how American households share national growth tied to the national account. At the end of this year, BEA’s personal income distribution will be published at the state level by adding the period of 2016-2018.

The annual Federal Partners offered valuable insight from agency leaders and provided exclusive insights on agency and data updates, particularly during a time when researchers are eager for new and real-time data.