The prospects for a revival in productivity may seem bleak. After all, there has been a lot of technological fatalism in the last decade: In 2013, Peter Thiel, a venture capitalist, pondered the technological advances of the moment that “we wanted flying cars, instead 140 characters”. Robert Gordon of Northwestern University has repeated this sentiment, speculating that humanity could never again invent something as transformative as the flush toilet. Throughout the decade, productivity data largely supported the views of the pessimists.
In addition, some studies of past pandemics and analyzes of the economic impact of this pandemic suggest that Covid-19 is likely to worsen the prevailing productivity picture. According to the World Bank, countries affected by pandemic outbreaks in the 21st century (excluding Covid) saw labor productivity declines by 9% compared to unaffected countries after three years.
And yet stranger things have happened. The brutal years of the 1930s were followed by the most extraordinary economic boom in history. A generation ago, economists had all but given up hope of ever achieving postwar performance when a computer-controlled productivity explosion took place. And today there is again tantalizing evidence that the economic and social traumas of the first two decades of this century may soon give way to a new period of economic dynamism.
Productivity is the magic elixir of economic growth. While an increase in the labor force or the capital stock can increase output, the impact of such contributions will decrease unless better ways of using these resources are found. Productivity growth, where more output is obtained from available resources, is the ultimate source of long-term income growth. It’s not everything, as a Nobel Prize winner Paul Krugman once remarked, but in the long run it is almost everything. However, economists know less about how to increase productivity than they would like. Increases in labor productivity (ie higher performance per worker per hour) appear to be due to improvements in education levels, higher investments (which increase capital per worker) and the introduction of new innovations. An increase in total factor productivity – or the efficiency with which an economy uses its productive inputs – may require the discovery of new ways to produce goods and services or the reallocation of scarce resources from businesses and places of low productivity to places of high productivity.
Worldwide productivity growth slowed sharply in the 1970s from the searing high rates in the post-war decades. From the mid-1990s to the early 2000s, there was an outbreak of higher productivity growth in the rich world, led by America. Emerging economies also experienced rapid productivity growth in the decade leading up to the global financial crisis, fueled by heavy investment and trade expansion, bringing more sophisticated techniques and technologies to developing world participants in global supply chains. However, since the crisis, there has been a broad and persistent slowdown in productivity growth (see Figure 1). According to the World Bank, around 70% of the world economy is affected.
Taking into account the slowdown is a difficult process. The World Bank believes that a slowdown in trade growth and fewer opportunities to adopt and adapt new technologies from richer countries may have contributed to depressing productivity gains in emerging economies. In all economies, sluggish investment is a culprit after the global financial crisis: a particular problem in places with an aging and shrinking workforce. While those headwinds certainly play a role, the bigger question is why seemingly powerful new technologies – like improved robotics, cloud computing, and artificial intelligence – haven’t resulted in more investment and higher productivity growth.
By and large, three hypotheses compete to explain this lull. One voiced by the technopessimists insists that, for all the enthusiasm for world-changing technologies, the latest innovations are simply not as transformative as the optimists insist. While it is possible that this will turn out to be correct, continued technological advancement makes it less and less plausible as an explanation for lull. AI may not have changed the global economy at the dramatically disruptive pace expected five to ten years ago, but it has become significantly, and in some cases, startlingly more powerful. GPT-3, a speech prediction model developed by OpenAI, a research firm, has demonstrated remarkable ability to hold conversations, draft long texts, and write code in surprisingly human ways.
Though the internet’s potential to support an economy where the limitations of distance are long undervalued, cloud computing and video conferencing have proven to be economically valuable over the past year, leaving large amounts of productive activity barely interrupted despite the formwork many offices. New technologies can clearly do more than has been generally required of them in recent years.
This suggests a second explanation for slow productivity growth: chronically weak demand. In this view, most voiced by Harvard University’s Larry Summers, governments’ inability to spend enough limits investment and growth. More public investment is needed to realize the potential of the economy. Chronically low interest and inflation rates, weak private investment and weak wage growth since the turn of the millennium clearly show that demand has been largely insufficient over the past two decades. Whether this will significantly undercut productivity growth is difficult to say. In the years leading up to the pandemic, as unemployment fell and wage growth rose, American labor productivity growth seemed to accelerate from an annual increase of just 0.3% in 2016 to a 1.7% increase in 2019: the fastest growth rate since 2010.
A third explanation, however, is the strongest argument for optimism: It takes time to figure out how powerful new technologies can be used effectively. AI is an example of what economists refer to as “general purpose technology” like electricity, which has the potential to increase productivity in many industries. However, making the best use of such technologies requires time and experimentation. This accumulation of know-how is really an investment in “intangible capital”. Recent work by Erik Brynjolfsson and Daniel Rock at MIT and Chad Syverson at the University of Chicago argue that this pattern leads to what they refer to as the “productivity J-curve.” When new technologies are first introduced, companies shift their resources to investing in intangible assets: developing new business processes. This shift in resources causes business performance to suffer in a way that cannot be fully explained by shifts in the measured use of labor and material capital and is therefore interpreted as a decline in productivity growth. Later, when intangible investments bear fruit, measured productivity increases because production increases in a way that cannot be explained by the measured work and material resources.
As early as 2010, the failure to consider intangible investments in software had little impact on productivity numbers, according to the authors. However, productivity was increasingly underestimated. By the end of 2016, productivity growth should be 0.9 percentage points above official estimates.
This pattern has occurred before. In 1987, Robert Solow, another Nobel Prize winner, noticed that computers were everywhere except in productivity statistics. Nine years later, American productivity growth began a decade of acceleration reminiscent of the golden economic age of the 1950s and 1960s. These processes are not always sexy. In the late 1990s, the rising share prices of flashy internet startups made most of the headlines. The result of the productivity growth had other causes, such as improvements in advanced manufacturing techniques, better inventory management, and rationalization of logistics and production processes made possible by digitizing company records and using clever software.
The J-curve provides a way to balance technical optimism and the introduction of new technologies with lousy productivity stats. The role of intangibles in unlocking the potential of new technologies can also mean that, despite its economic damage, the pandemic has increased the likelihood of a productivity boom developing. Office closings have forced companies to invest in digitization and automation or to make better use of existing investments. Old analogous habits could no longer be tolerated. While no business statistics show this, in 2020 executives around the world invested in the organizational overhauls necessary for new technology to work effectively (see Figure 2). Not all of these efforts will have resulted in productivity improvements from the pre-pandemic norm. With the withdrawal of covid-19, the companies that have changed their operations will maintain and build on their new approaches.
The crisis forced changes
Early evidence suggests that some transformations are very likely to continue and that the pandemic has accelerated the pace of technology adoption. A survey of global companies conducted by the World Economic Forum earlier this year found that more than 80% of employers intend to accelerate plans to digitize their processes and offer more remote working opportunities, while 50% intend to accelerate the automation of production tasks. Around 43% expect changes like this to result in a net decrease in their workforce: a development that could pose challenges in the labor market but, by definition, implies improvements in productivity.
More difficult to assess, but no less realistic, is the possibility that moving so much work to the cloud could have productivity-enhancing effects on a national or global level. High housing and real estate costs in rich, productive cities have excluded businesses and workers from places where they may have accomplished more with fewer resources. When technicians can more easily contribute to top businesses while living in affordable cities off the American coast, strict zoning rules in the Bay Area of California will become less of a bottleneck. Office space in San Francisco or London that is freed up by the increase in remote working could then be occupied by companies that really need their employees to work in close physical proximity. In addition, and if policy allows, the pandemic-induced surge in distance learning and telemedicine could help usher in a new growth phase in trade in services and create economies of scale in sectors that have long proven resistant to productivity-enhancing measures.
None of this is taken for granted. To make the most of new private investments in technology and know-how, governments need to bring about rapid recovery in demand, make complementary investments in public goods like broadband, and focus on filling the educational gaps that have plagued so many students have a consequence of school closings. But the raw materials for a new productivity boom seem to fit in a way that has not been seen in at least two decades. The darkness of this year could actually mean that dawn is close to the horizon.
