Global Economics Comment: Technology and the Productivity Rebound (Zhestkova)
Table of Contents
Labor productivity growth disappointed in the decade preceding the pandemic, averaging only ½ -1% in the US, the Euro Area and the UK. We have argued previously that some of this weakness may reflect mismeasurement of the positive impact of new technology. But even leaving aside this issue, we see three signs why the underlying pace of measured productivity growth may now be accelerating.
The first is mean reversion. Growth in both labor productivity and total factor productivity (TFP) appears to have picked up from rock-bottom levels in the later part of the past decade in the US, the UK, and the Euro Area. This is consistent with our finding that TFP growth historically tends to mean-revert slowly over time and suggests that further productivity gains are ahead.
The second is evidence of recent technological acceleration, and potentially the second wave of the IT revolution. Investments in intellectual property products (IPP) had already started to rise before 2020 in the major advanced economies. Looking across US sectors, rapid productivity growth in tech-related service sectors that invest heavily in IPP has driven much of the pickup in US productivity growth since 2016.
The third is evidence of increased economic dynamism. New business formation and US patent applications have surged since the start of the pandemic, and the measured contribution of the IT sector to US productivity growth has picked up significantly in the recent years.
Technology and the Productivity Rebound
Labor productivity growth disappointed in the decade preceding the pandemic, averaging only ½ -1% in the US, the Euro Area, and the UK (Exhibit 1). We have argued previously that some of this weakness may reflect increasing mismeasurement of the positive impact of new technology.[1] Leaving aside this issue, we first revisit key explanations of the slowdown, then analyze the latest evidence for these explanations, and conclude that the underlying pace of measured productivity growth may now be accelerating.
Exhibit 1: Soft Productivity Growth in the Past Decade
Source: Haver Analytics, Goldman Sachs Global Investment Research
Economic studies have offered three interrelated explanations for the productivity slowdown after the GFC (Exhibit 2). The first is mean reversion of total factor productivity growth (TFP) — a residual measure of technological progress — and therefore a slowdown in productivity growth after the boom of the late 90s and early 2000s. The second is that it takes a substantial amount of time before most businesses implement recent technological breakthroughs (for instance, in AI). The third is a potential decline in economic dynamism. Revisiting the latest evidence for these three factors, we see signs that the underlying pace of measured productivity growth is currently on an upward trajectory.
Exhibit 2: Mean Reversion After the Early 2000s Boom, Technology Implementation Lags, And Declining Dynamism Help Explain the Post-GFC Productivity Slowdown
Source: Goldman Sachs Global Investment Research
Sign #1: TFP Mean Reversion
Growth in both labor productivity and total factor productivity (TFP) appears to have picked up from rock-bottom levels in the later part of the past decade in the US, the UK, and the Euro Area (Exhibit 1). This is consistent with our earlier statistical finding in a large postwar panel of advanced economies that TFP growth historically tends to mean-revert slowly over time.[2]
Over a longer time horizon (going back 140 years), US TFP growth appears stationary, with a variety of cycles of strength and weakness around an average annual growth rate of 1.2% (Exhibit 3). This implies that a decline in TFP growth to rock-bottom levels is likely to be followed by a recovery, and suggests that a further acceleration in productivity growth to the levels consistent with its historical mean is likely ahead.
Exhibit 3: TFP Growth Tends to Slowly Mean Revert
Source: Kendrick (1961), World Penn Tables, The Conference Board, Federal Reserve Bank of San Francisco, Goldman Sachs Global Investment Research
Sign #2: Technological Acceleration Started Before the Pandemic
Technological waves are often the economic driver of the statistical tendency for TFP growth to return to its long-run average. There is some evidence that technological acceleration, and potentially the second wave of the IT revolution, have contributed to the recent pickup in productivity growth. Investments in intellectual property products (IPP) had already started to rise before 2020 in the UK, the Euro Area, and especially the US, and surged in the first year of the pandemic (Exhibit 4).
Exhibit 4: A Pickup in Intellectual Property Products Investments
Source: Haver Analytics, Goldman Sachs Global Investment Research
Interestingly, the pick-up in US productivity growth started around the same time, about 5 years ago, as this acceleration in IPP investments. Moreover, labor productivity has grown particularly quickly over this period in tech-related service sectors that invest heavily in IPP, such as IT services and broadcasting and telecommunication (Exhibit 5).
Exhibit 5: A Positive Relationship Between IPP Investment and Productivity Growth Across US Industries
Source: Haver Analytics, Goldman Sachs Global Investment Research
While the manufacturing sector drove much of aggregate US productivity growth in 1996-2010[3], tech-related services have driven much of the acceleration over the last five years (Exhibit 6).
Exhibit 6: The Contribution from Tech-Related Services to US Productivity Growth Has Been Rising
Source: US Bureau of Labor Statistics, Haver Analytics, Goldman Sachs Global Investment Research
Sign #3: Pandemic Has Jolted Business Dynamism and Innovations
The third argument for a persistent acceleration in labor productivity growth is evidence of increased economic dynamism. Exhibit 7 shows that new business applications surged in the US in mid-2020 and remain unusually high.[4] France and Germany also experienced a significant but less dramatic pick-up in new business formation. Economic research suggests that a rise in entrepreneurship should boost productivity growth as new firms are an important source of innovation-driven TFP growth.[5]
Exhibit 7: A Rise in New Business Formation
Source: Haver Analytics, Goldman Sachs Global Investment Research
Similarly, US patent applications have also risen sharply over the last year (Exhibit 8). The rise in applications has been particularly pronounced for patents in the “electricity” (e.g. electric equipment, broadcasting, wireless network) and “physics” (e.g. information and communication technology, computations) groups, which often cover fundamental inventions with longer-run productivity gains, and the usually more applied innovations in the “human necessities” (e.g. food, household appliances, entertainment) group with quicker productivity gains.
Exhibit 8: A Pandemic Jump in US Patent Applications
The series are adjusted for truncation, resulting from a gap between an actual date of application filing and the date when the application is published on USPTO website after initial processing.
Source: US Patent and Trademark Office, Goldman Sachs Global Investment Research
While these surges in entrepreneurship and patents bode well for future technology-driven productivity growth, the contribution from IT sectors — defined as computer hardware, software, communication equipment, semiconductors, and R&D — to US labor productivity has already surged last year (Exhibit 9), with a particularly large contribution from a greater use of IT capital (“capital deepening”).[6]
Exhibit 9: A Pandemic Jump in the Contribution of the IT Sector to US Productivity Growth
Source: US Bureau of Labor Statistics, Semiconductor Industry Association, Haver Analytics, Goldman Sachs Global Investment Research
Yulia Zhestkova
Appendix
Exhibit A: Decomposition of Labor Productivity Growth in the US
Source: US Bureau of Labor Statistics, Semiconductor Industry Association, Haver Analytics, Goldman Sachs Global Investment Research
- 1 ^ We estimated that the pace of annual real GDP growth is understated by around 1.0pp. This captures unmeasured nominal output (e.g. tax-haven profit shifting), the value of free digital goods (e.g. social media, search engines), business investment in new types of intangible ICT capital, and the overstatement of consumer inflation (e.g. quality change, outlet bias). See Spencer Hill, “Productivity Paradox v2.0: The Price of Free Goods”, US Economics Analyst, 14 July 2019.
- 2 ^ See Zach Pandl and Daan Struyven, “Historical Context on the Productivity Slump”, US Economics Analyst, 22 April 2016.
- 3 ^ Note that computers and peripheral hardware equipment fall into manufacturing sector.
- 4 ^ The increase in new business applications in the US is fairly broad-based across sectors. For more details on US business dynamism during the pandemic, see Joseph Briggs, "Business Sector Scarring Effects: Damage to the Supply Side Still Looks Modest", US Daily, 24 June 2021.
- 5 ^ See Daron Acemoglu, Ufuk Akcigit, Harun Alp, Nicholas Bloom, and William Kerr. 2018. “Innovation, Reallocation, and Growth”, American Economic Review, 108(101): 3450-3491
- 6 ^ To estimate the net contribution of IT sectors, we follow Byrne, Oliner, and Sichel (2013) and use a variety of GDP-accounting methods to decompose output per hour growth into the contributions of labor quality, capital deepening and TFP (net of utilization and quality adjustments as estimated by the San Francisco Fed). We then distill the contribution of capital deepening and TFP coming from IT sectors using micro data on sector-specific inputs in production. Exhibit A in the Appendix shows the result of this decomposition including the contribution of non-IT related components.
Investors should consider this report as only a single factor in making their investment decision. For Reg AC certification and other important disclosures, see the Disclosure Appendix, or go to www.gs.com/research/hedge.html.