What’s happening to the growth of productive new firms?
One of the potential drivers of productivity growth is reallocation of labor and capital. Workers moving to new jobs, unproductive firms disappearing, and new firms appearing are supposed to help move resources to their best uses. The misallocation of labor seems to be a big factor underlying the tepid productivity growth in the United States over the past decade. But it also seems that some of the sources of reallocation that once helped boost productivity aren’t working as well as they did in the past.
The new work comes from a number of researchers who together and in smaller groups have documented the decline in the startup rate in the United States, as well as the slowing pace of growth for fast-growing firms.
In their new paper, Ryan Decker of the Federal Reserve Board, John Haltiwanger of the University of Maryland, and Ron Jarmin and Javier Miranda of the U.S. Census Bureau aim to understand what might be causing the decline in business dynamism—the rate at which new businesses are started and destroyed. They look specifically at the trends within the high-tech manufacturing sector, a section of the economy that has micro-level productivity data and experienced declining dynamism over the 2000s.
The authors point to two potential sources of declining dynamism. First, the kinds of productivity shocks that firms experience—such as figuring out a new technique that makes workers more productive—could have changed over time. The authors refer to this as a changing volatility in productivity shocks, which in the context of declining dynamism would mean that firms are experiencing fewer positive productivity shocks. But the economists rule that out once they look at the firm-level productivity data.
The other possibility is that firms are less responsive to the productivity shocks they experience. In other words, a firm may become more productive but its employment growth will be slower than it was in the past. More productive firms, at least in the high-tech manufacturing sector, are contributing fewer jobs than they were in the past. Furthermore, this trend holds up for both young firms and mature firms. If labor reallocation depends on the most productive firms growing faster to provide new more productive jobs, that mechanism doesn’t seem to be working as well post-2000.
This analysis is clearly limited in that it only looks at one sector of the economy, in part due to data limitations. As the authors point out, however, the changing patterns of dynamism for the overall high-tech sector look very similar to the trends for high-tech manufacturing. (If only there were good establishment-level data on total factor productivity for social media companies.)
But the declining productivity growth in a high-tech sector does seem to align with work by John Fernald of the Federal Reserve Bank of San Francisco, who has argued that the productivity decline of the last decade has been concentrated among IT firms and firms that use IT intensively. A provocative thesis to say the least, but given the evidence presented in this paper, it’s something to consider even more.