Measuring Tech Remains Tricky and Likely Undercounts Productivity, Economists Say

Increased stakeholder cooperation and revisiting historic assumptions on prices, quality and quantity are needed, panelists said when we asked about how agencies can exchange data among themselves and with the private sector. Some such efforts incorporate Amazon, said Conference Board Economics Program Research Director Carol Corrado. Historically in the U.S., stats on price and quantity were collected differently, said Stanford professor of economics Susan Athey. "Still do," interjected Corrado. That’s a problem when consumers buy things on sale.

All agencies have initiatives to "exploit private data as much as possible" and often work with firms "to help speed up that process," Corrado said. "Private data can be wonderful at augmenting existing statistics to tell you more about what’s really going on at the regional or micro level. But it can't supplement the broad surveys that sort of ground a lot of the basic statistics. And that’s even true for administrative data from government sources." Agencies may need to update how they structure and/or do surveys, economists said. It's "a huge task for the economics community to wrap its head around," said Leonard Nakamura, Philadelphia Federal Reserve Bank emeritus economist. "We need a major cooperative effort between the economics community and the statistical agencies." He sought a "major increase in the statistical agencies' budgets."

There's been a slowdown in productivity growth since around 1948, said moderator and TPI President Scott Wallsten (see his slide). "Why, with all this technological innovation, do we see productivity growth still so low," he asked. He noted 2018 had growth of about 1 percent, an uptick.

Nakamura thinks "the entire slowdown is due to mismeasurement." The economy "has accelerated, rather than slowed," he said. "We’ve had a very dramatic slowdown in real GDP growth" even amid self-driving cars, a wide array of newer tech, DNA sequencing and the like, he said. "We don’t measure the kind of economic growth that we’ve been creating." And "as we measure them, they contribute negatively to real" gross domestic product, he said. "GDP does not do well with zero pricing" and some tech products are free, he noted. Nakamura's research doesn't necessarily reflect the Philadelphia Fed's views, a bank spokesperson said when we asked if the Fed bank endorsed his conclusions.

Agencies have had some stumbles. Some agencies realized "they just totally dropped the ball" on measuring the digital economy, said Corrado. For the Bureau of Labor Statistics, Uber drivers may not fully report their employment, so stats don’t pick this up, she noted. BLS went some time without a leader, which might have slowed improvements, but things should improve "before too long," she predicted. A BLS spokesperson pointed us to a 2017 academic-journal article by agency economists on how government stats change to account for new digital-age goods.

The U.S. Bureau of Economic Analysis and the Census Bureau are working on improvements, too, the panel was told. The Census Bureau didn't comment. "BEA is constantly looking for ways to better measure the fast-changing U.S. economy," emailed a spokesperson. "Many categories of high-tech categories of goods experience rapid increases in processing speed, memory and other capabilities that are not captured by traditional price measures. We have for years been at the forefront of developing price measures that capture these improvements in quality."