A recent paper by John Fernald, a senior researcher for the Federal Reserve Bank of San Francisco, titled ‘What is the New Normal for US Growth?’ suggests that for the US economy it could be in the range of 1.5-1.75 percent (in real terms). The author’s preferred point estimate is 1.6 percent which he sees as being ‘consistent with productivity growth net of labor quality returning over the coming decade to its average pace from 1973–95, which is a bit faster than its pace since 2004.’
The paper notes that productivity growth averaged 2.5-2.75 percent in its better periods (e.g. 1948-1973 and 1995-2004) and in the slower periods (e.g. 1973-1995) it was as low as 1.25 percent. But in the 2010-2015 period it has been only ~0.25 percent per annum on average. The current period of slower economic growth is mainly attributed to demographics and educational attainment: ‘As baby boomers retire, employment growth shrinks. And educational attainment of the workforce has plateaued, reducing its contribution to productivity growth through labor quality.’
GDP growth is the sum of growth in hours worked and GDP per hour. Fernald notes labour force growth shows significant variation over time: ‘in the 1970s and 1980s, the labor force grew much more rapidly than the population as the baby boom generation reached working age and as female labor force participation rose. Those drivers of labor force growth largely subsided by the early 1990s.’ GDP grew at nearly 3% in the 1973-95 period but if hours worked had only grown at 0.5 percent (the current Congressional Budget Office projection for the next decade) and productivity grown at its historic rate then GDP growth would have been closer to 1.75%.
While the slowdown from demographic changes seems clear the productivity effect is less so: predicting the next technological leap is a lot more difficult. The more pessimistic camp would argue technological leaps forward such as the steam engine, electric dynamo, internal combustion engine and microprocessor that were key drivers of productivity gains in the past are unlikely to be easily repeated. But at the more bullish end of the scale, Accenture released a study on 12 developed economies, including the US, arguing that artificial intelligence (AI) could double annual economic growth rates: they argue ‘AI has the potential to boost labor productivity by up to 40 percent’ by automating a lot of tasks and freeing up workers to do more skilled tasks. But we note that this will also eliminate a lot of jobs. Paul Daugherty, one of the report’s authors, sees this as a ‘short-term displacement’ but as research by Forrester notes within 5 years AI could cut 6 percent (net figure) of US jobs, a huge number, so we are less convinced of a massive productivity boost in the next 5 years or so.
The Fed’s own updated GDP projections from September seem consistent with a lower normal: the median forecast for 2016 GDP growth is 1.8 percent, edging up to 2 percent in 2017 and 2018 and back down to 1.8 percent in 2019. Importantly, the median estimate of the long-run normal rate of GDP growth was cut to 1.8 percent from 2 percent in June.
For us, the San Francisco Fed paper just reinforces our view that we are in an era of low growth and the neutral rate of interest will remain low underpinning interest rate expectations at the long end of the curve. Thus, the hunt for yield seems set to continue and for us the positive yields offered on US dollar denominated debt from creditor nations looks to be a compelling opportunity.