Last week I was fortunate to attend the Federal Reserve Bank of Kansas City’s Economic Policy Symposium in Jackson Hole, Wyoming, which brought together policymakers, economists, members of the media and other experts. This year’s symposium emphasized the implications of changing market structure, the growth of online retail, and the evolution of financial markets for the economy. As someone from a statistical agency, I was struck by how much the data employed in discussions like this has evolved over the years, especially the shift from relying on macroeconomic indicators largely built from sample-based surveys to statistics built from very granular data on workers, firms or even transactions. 

I have spent much of my career researching and developing just these sorts of statistics from the rich microdata that underlie the U.S. Census Bureau’s primary data products, and work along these lines continues at statistical agencies around the world. Another important aspect of the data discussed in Jackson Hole is the shift to new, mostly private, data sources. This trend is just getting underway and will be a primary driving force behind improvements in economic measurement going forward.

These efforts rely on long-term investments in research and infrastructure and a foundation of collaboration between the agencies, academia and business. A long history of collaboration between federal statistical agencies underlies many of the principal federal economic indicators that the participants in Jackson Hole and many others use to gauge the health of our economy, including gross domestic product, price indexes and the unemployment rate.

The Conference on Research in Income and Wealth (CRIW) is a key forum for collaboration that dates from 1936, when the National Income and Product Accounts were first being developed. Administered by the National Bureau of Economic Research, the CRIW’s current institutional partners include the Board of Governors of the Federal Reserve System, the Census Bureau, the Bureau of Economic Analysis (BEA), the Bureau of Labor Statistics (BLS), the Statistics of Income (SOI), and Statistics Canada. The rich cooperation between academic economists and government statisticians has produced many of the key improvements in economic measurement over the past 82 years. This work continues with Big Data for 21st Century Economic Measurement, a new conference organized by former BLS Commissioner Katherine Abraham, BEA Director Brian Moyer, University of Michigan Professor Matthew Shapiro, and me.

This upcoming conference, planned for March 2019, will examine how the explosion of digital data from online transactions, sensors, and so on, which reside mostly in the private sector, can be repurposed to provide more timely and granular economic statistics. Increased involvement from the private sector in systematic economic measurement is a welcome trend and certainly not limited to providing data to researchers from academia, the Federal Reserve Bank and the statistical agencies. Indeed, several companies have begun producing publicly available economic statistics based on their corporate data assets. Examples include the JPMorgan Chase Institute, which publishes a range of reports based on summaries of bank account data, and Automatic Data Processing, which produces payroll jobs estimates that complement data from BLS. The National Association of Business Economics, a professional association for business economists, has taken a leadership role in convening private-sector economists and data scientists who are exploring ways to use corporate data assets, not only to drive value for their companies, but to build economic statistics for public use.

While there is justified excitement about these new data sources, which I share, the data also have important limitations that can make it difficult to use them to create official statistics. It is essential for the Census Bureau, BEA and other federal statistical agencies to research carefully how best to use these data. A first step which is already underway involves linking them to existing administrative, survey and census data assets to better understand their strengths and weaknesses for tracking economic activity.

Increased collaboration between the statistical agencies and private-sector data providers is essential to this research so that federal statistical data can continue to be trusted by policymakers and researchers, like those featured at this year’s Jackson Hole symposium.