The key call in any macroeconomic forecast is judging how fast the real output of an economy will grow over the longer term. Trend growth, or the expansion rate of potential output, has three parts: the pace of the increase in population, their participation in the workforce, and the growth rate of output per hour worked. The first two components are about demographics, and demographics is mostly preordained in forecasting since its forces typically evolve slowly.
The third part, output per hour or productivity, is more challenging.1 As seen in the first chart, the quarter-to-quarter gyrations of productivity growth in the United State are considerable. In fact, in about 1980, economists belatedly identified markedly slower US productivity growth starting in 1973. As a result, there was a nearly decade worth of mistakes at the Federal Reserve, when policymakers thought that the economy had more room to run than was actually the case.
The solid line smooths through the quarterly changes in output per hour by taking a five-year moving average, which shows the low frequency or trend movements. Herein lies the bad news about the economic outlook: the growth of productivity has slowed significantly, starting from a local peak around 2003, so that potential real GDP accordingly tracks along a reduced path.
Any economic forecast, including ours, gravitates toward that pole. Despite ongoing financial accommodation and significant fiscal impetus, US real GDP growth is only in the upper 2 percent range in 2018 and falls back next year. Even so, this tepid performance of aggregate demand, by historical standards, imparts pressure on resources and costs.
1 Chad Syverson of the University of Chicago’s Booth School presented an overview of these issues at the Federal Reserve Bank of Atlanta’s recent annual financial markets conference, found here: https://www.frbatlanta.org/-/media/documents/news/conferences/2018/0506-financial-markets-conference/papers/syverson-chad-ai-productivity-paradox-for-distribution.pdf.