Robert Gordon and the case for supply-side secular stagnation and techno-pessimism
Advanced democracies are in a period of stagnation reflecting a loss of capitalist dynamism and a failure of new technologies to generate enough growth. That’s a problem for social democrats.
This Reviving Social Democracy newsletter is premised on the idea that social democrats must change their message to reflect the reality of secular stagnation and the limits of technological solutions (techno-pessimism). Government must invest more directly into the living standards of working and middle-class families.
Summary
Productivity growth is highly correlated with improved living standards. Unfortunately, we have not seen meaningful levels of productivity growth in the advanced democracies in the opening decades of the 21st Century resulting in stagnant real wages for the majority of middle-income workers and even declining real wages at lower income levels. Economic insecurity is rising and living standards are eroding.
There is reason to believe that we are stuck in a period defined by secular stagnation which has been made worse by aging populations, weak labor force participation, high levels of inequality, high levels of debt, and growing lower-wage service sectors.
Hopes that the information and communications technology (ICT) revolution and emerging technologies like artificial intelligence, robotics, and genomics would spur a new era of strong productivity growth and improved living standards remain unfulfilled.
This provides support for Robert Gordon’s secular stagnation theory, his belief that capitalist dynamism and potential output has reached near and medium-term limits.
Social democrats have no choice at this point but to ensure that governments reject passive acquiescence to markets and pursue direct government investment in projects and programs that directly impact the living standards and economic security of the working and middle classes.
[Note: Pages in parentheses below refer to Gordon’s The Rise and Fall of American Growth (2016), Princeton University Press. Other articles by Gordon are available at the links. For sources of other references and data, please leave a note in the comments and I’ll send them along. A 2013 12-minute TED video by Gordon can be viewed here. Although 10-years old, the video explains the essence of his thesis.]
Slow growth as the new normal
The idea that an advanced market economy might experience a period of no or slow growth lasting well beyond the normal business cycle first arose in the 1930s when the global Great Depression was extending deep into the decade. It was a controversial claim, dubbed “secular stagnation” by Harvard economist Alvin Hansen, and it quickly receded as New Deal spending in the U.S. and war spending everywhere catalyzed economic growth. The stimulus continued into the postwar period through the reconstruction of Europe and Japan, and the ensuing dynamic growth and rising living standards across the advanced economies put the matter to rest.
Most observers accept, however, that the postwar period of strong growth ran out of steam in the 1970s. Growth in U.S. labor productivity in the period 1972-1996 dropped by half compared to the rate of 1950-1972 and average real wages of low and middle-income workers everywhere started a 50-year period of stagnation.
At the outset, the problem appeared to be the OPEC oil shock and ensuing stagflation, a toxic combination of high unemployment and high inflation, but the trend continued even after inflation and unemployment dropped and oil prices stabilized. Since weak productivity growth was a key factor, though, political parties across the spectrum became convinced that the solution was to aggressively promote private sector business prospects. This included the so-called Third Way social-democratic governments of Clinton, Blair, and Schröder in the 1990s. Third Way policies aligned with neoliberal policies as corporations and financial institutions benefited from reduced regulation, permissive rules on financial risk, and corporation-friendly trade agreements. Welfare programs were cut back and budget deficits reduced. Business-friendly tax policies were promoted and industrial unions lost much of their bargaining power and political influence as capital moved relentlessly into low-wage regions and nations.
Neoliberal and Third Way policies were initially bolstered by a renewal of U.S. economic growth in the 1990s driven by rising investment in ICT starting in the 1980s. The period 1994 to 2004 in the U.S. saw meaningful gains in labor productivity as average output per hour and average real wages grew. Though Europe experienced less of a boost at the time, the new technology was rapidly penetrating key areas of the economy and Europe would start catching up.
The Great Recession, slow recovery, and the secular stagnation hypothesis
In retrospect, this strong economic growth based on ICT innovation only serves to draw attention to its limits. By 2004, stagnation had resumed most everywhere. Average real wages returned to the earlier trend exposing the often overlooked fact that the higher incomes of workers in the new ICT sector were being offset by declining or flat real wages for most others.
There was a parallel decline in defined-benefit pension plans in the U.S. while austerity and “reforms” were chipping away at pension trusts in other developed nations. Younger workers were especially vulnerable, being less likely to participate in the new 401K-type plans that replaced guaranteed pensions in the U.S. and more likely to be burdened with supporting a growing elderly population in all nations.
Weak economies also fostered rapid growth in two-earner working and middle-class families. Women entering the workforce is, of course, a welcome social effect, but while some women entered the workforce to pursue rewarding professional careers, most at that time took low-wage service jobs that were needed to maintain a middle-class lifestyle for their families. By contrast, in the 1950s and 60s, many semi-skilled (blue collar) single-earner heads of household in the U.S. could afford a car and a modest home in a suburban development. They could also afford to put a couple of kids through a public community or state college. They increasingly self-identified as part of a rising middle class.
The period since 2004, however, has been characterized by insufficient long-term investment in the private economy and real GDP per capita growth in the U.S. dropped by nearly half between 2000 and 2007. Too much capital began flowing into short-term and speculative investments that created the “dot com” boom and bust, and the boom and bust in real estate-based securities that triggered the global Great Recession in 2007.
At the outset of the Great Recession, most economists expected that the effects, though widespread and deeply felt, would still be within the scope of business-cycle dynamics, especially as governments injected massive amounts of liquidity into private economies worldwide to dull the worst effects. By 2020, however, a full twelve years later, growth rates had still not recovered to pre-recession levels despite declining unemployment, low inflation, low interest rates, and continued government efforts to boost private business.
In 2013, former U.S. Treasury Secretary, Larry Summers, worried that there was more going on than just a strong correction. He suggested that, as in the 1930s, the world’s advanced economies had entered a period of persistent slow growth with no clear endpoint and famously revived the idea of secular stagnation. Six years later, in 2019, Summers reiterated his analysis given weak responses to monetary stimulus and near zero interest rates in most developed nations. While unemployment steadily dropped during the recovery, wages barely budged.
A depression was probably avoided in 2007-8 through the massive injection of liquidity, but central bank monetary policy efforts were not able to shake economies out of their stupor. In 2016, Summers declared that central banks were running out of monetary policy ammunition and called instead for massive fiscal stimulus.
Gordon’s thesis: It’s more than weak demand; growth potential is also weak
Summers looks at stagnation through a demand-side lens. He sees the problem as a gap between potential growth and actual growth caused by a lack of demand. Yet he and others saw then and continue to see now the possibility of overcoming demand weakness through government spending. Summers does not accept what he calls the fatalistic view that the economies of the industrialized world are condemned to suffer permanent stagnation.
Economist and historian Robert Gordon is less sanguine about the ability of even extraordinary stimulus programs to reverse stagnation. His 2016 The Rise and Fall of American Growth paints a more worrisome and fatalistic picture of what has been happening since the 1970s, a point reflected in the title of his 2016 article American growth has slowed down. Get used to it. While Gordon’s book develops his thesis mostly in relation to the U.S., he sees it affecting the entire advanced sector since productivity has been declining at more or less the same rate everywhere since 2004.
Gordon believes economies will adapt and find ways to muddle through going forward with intermittent growth occurring in specific industries and nations. At the end of the day, however, he identifies the central longer-term problem, a lack of capitalist dynamism and the limited potential of the next generation of technological innovation.
Gordon starts his 2016 book by pointing out that labor productivity and economic growth were stuck at low levels throughout human history. He cites a calculation showing that the annual rate of growth in the Western world was a mere .06% percent per year between AD 1 and AD 1820, 6% per century, representing roughly the modern equivalent of living on $400-$600 per year, just above subsistence (pp. 2-3).
Gordon (pp.1-6) then goes on to identify a flood of inventions and technological innovations in the mid-19th to mid-20th Centuries, which he calls the “Great Inventions”, from electricity and indoor plumbing to the internal combustion engine, refrigeration, and electronics, that radically transformed the more advanced economies and societies in just a few decades. Productivity and GDP grew rapidly and living standards increased significantly. He provides numerous, often dramatic, examples of the stunning changes that defined what he calls the “Special Century,” roughly 1870 to 1970. It was responsible for (p.1):
…freeing households from an unremitting daily grind of painful manual labor, household drudgery, darkness, isolation, and early death. Only one hundred years later, daily life had changed beyond recognition. Manual outdoor jobs were replaced by work in air-conditioned environments, housework was increasingly performed by electric appliances, darkness was replaced by light, and isolation was replaced not just by travel, but also by color television images, bringing the world into the living room. Most important, a newborn infant could expect to live, not to age forty-five but to age seventy-two.
Gordon did not, however, see this as a continuing trend.
The economic revolution of 1870 to 1970 was unique in human history, unrepeatable, because so many of its achievements could only happen once.
The effects of the Special Century were broad-based. Wages rose and relative prices of essentials dropped. People moved from farms to cities and created a new form of social existence, suburbs. Automobiles and trucks became ubiquitous and highways soon criss-crossed all the advanced nations. Electricity began permeating every aspect of life and could be easily and cheaply brought out beyond the inner city. A massive middle class was emerging and demand for higher-skilled labor rose along with demand for skilled professionals. The land grant university systems were rapidly expanding as regional and community colleges were established in all states and with curriculums that included science, engineering, technology, and the practical arts. Increased leisure time and disposable income gave birth to major new industries, from travel and hospitality to fashion, sports, and entertainment.
Europe and Japan were part of this transformation, made easier by an already existing literate workforce and a mature science and engineering research and development capacity.
As with Summers, Gordon sees the 1970s as the inflection point for a shift into secular stagnation. Differing from Summers, however, he suggests that things changed not just because policymakers failed to anticipate and combat declining demand, but because the impact of the Great Inventions had run its course and was already baked into the numbers. The problem, he suggests, was not just a gap between potential growth and actual growth, but rather a decline in potential growth. Further growth would be modest and vulnerable to exogenous effects.
Gordon (p. 2) notes that what economists call Total Factor Productivity grew in the U.S. after 1970 at only a third of the rate achieved between 1920 and 1970 as the productivity gains spurred by the Great Inventions were reaching their natural limits. Today’s cars are safer and have more features, but the improvements are minor when compared to the revolution created when affordable versions rolled off mass production lines in the 1920s and transformed the world. Except for the addition of microwave ovens, kitchens today look and work much like kitchens in the 1960s. Food yields from an acre of land cannot grow much greater than the dramatic increases that resulted from the earlier technologization of agriculture, and there are far fewer opportunities to further reduce the size of the farm workforce. Increasing lifespan beyond 70 years of age adds little to labor productivity and creates new social costs.
Gordon provides many more examples of why he feels that further innovation in the major non-ICT domains has intrinsic, even biological, limits and that further major increases in labor productivity and living standards will be hard to come by. Societies should try to push forward but would also need to adapt.
Gordon also brings up several important exogenous factors that have further inhibited growth since the 1970s, which he calls “headwinds”. Like many economists, he worries about rising inequality (pp. 608-620) and slow labor force growth in the face of aging societies (pp. 627-629). He is also concerned about large and growing debt to GDP ratios (also, pp. 629-30), along with a factor specific to the U.S., education and training deficits caused by budget cuts, tuition increases, and poor outcomes at the K-12 level (also, pp. 624-627). He sees this as compromising potential growth in higher-skill areas of the U.S. economy.
According to Gordon (pp. 634-639), such headwinds are dragging real growth rates down even further, to below 1% in most advanced nations. He predicts that, through 2040, disposable median income per person in the U.S., his proxy for the material standard of living, will grow by only .3% per year, less than a fifth of the rate achieved between 1970 and 2014 and a seventh of the rate between 1920 and 2014 (p. 637). He sees the situation as no better, on average, in Europe, Japan, and OECD nations generally.
The limited economic impact of the ICT revolution
Many have responded to Gordon by invoking what they see as the continuing transformative power of the ICT revolution. Techno-optimists suggest that economic numbers do not adequately reflect the new reality and that we have only just begun to benefit from the efficiencies and productive potential of computer and telecommunications innovations, especially the internet. They also point to the many ways that the new technologies have contributed to general quality of life, something not easily quantified.
Gordon is skeptical. He agrees that the new digital technologies have provided a boost in productivity, economic growth, and real wages. He also agrees that it has added a perhaps unmeasurable qualitative dimension to our lives. But he disagrees about the scale of the impact and the longer-term potential. He considers the boost to be both narrowly defined and of limited duration, in what he calls
“the narrow sphere of human activity having to do with entertainment, communications, and the collection and processing of information.” (p.2)
Most real wage growth, for instance, has been associated with workers and professionals in the ICT sector. He contrasts this with the slowdown underway in other major spheres of economic activity (p. 2):
“For the rest of what humans care about – food, clothing, shelter, transportation, health, and working conditions both inside and outside the home – progress slowed down after 1970, both qualitatively and quantitatively.” (p.2)
Gordon (pp. 577-584) believes that most ICT efficiency gains had already been accounted for by 2015. Bank tellers, secretaries, and telephone operators have been gone for decades and the majority of manufacturing processes that can be handled by computer-controlled robots are already being so handled while those requiring hand assembly have moved to low-wage nations. It has been 30 years since the invention of the barcode, but humans still stock the shelves and dominate the check-out lines at supermarkets.
Telephone landlines disappeared quickly and smart phones and other personal electronic devices became ubiquitous many years ago along with HD TVs and laptop computers. New generations of upgrades of electronic devices have added only marginally to their capabilities, though, causing increasing numbers to delay upgrades until forced by the manufacturers. Entertainment and social content, a large component of the recent surge in technology-based enterprise, have proliferated so extensively that it has become nearly impossible for an individual to absorb any more. Despite the fantastic ways information technology has added convenience and entertainment to our lives, its overall economic impact has been relatively subdued.
Gordon presents a sobering picture and he sees little to be gained from newly emerging technologies like artificial intelligence, big data analytics, 3-D printing, gene therapy, and next-generation robotics, including driverless cars. Robotics for manufacturing has already reduced the manufacturing workforce substantially and 3-D printing will likely not go much beyond the production of prototypes for the foreseeable future. As already noted, most gains in lifespan and health have reached their biological limits and genomics and next generation medical technology will provide only modest additional benefits. Extending the human lifespan much further will result in many more living with disabilities in their 80s and 90s when dementia, mobility disorders, and frailty are at peak prevalence.
Artificial intelligence holds great potential, particularly in scientific research and areas like medical diagnostics, but the technology will be controlled by private companies and, like social media companies, will primarily be driven by a search for profits.
While techno-optimists have challenged Gordon’s pessimism about the ICT future, they are largely guessing as to how much these emerging technologies will add to labor productivity and economic growth. Andrew McAfee and Erik Brynjolfsson suggest that the productivity impacts of artificial intelligence, big data, and internet platforms are just around the corner and will be transformative, but Benedikt Frey and Michael Osborne at Oxford estimated in 2013 that as many as 47% of all jobs in the U.S. would be done by machines in “perhaps a decade or two”. It’s been a decade since that prediction and it’s already proven to be a gross overestimation.
While it’s difficult to predict the future, there are reasons to doubt the techno-optimist vision. In most cases to date, these new technologies are being used alongside human employees and sizable net job loss due to next-generation technology have yet to be seen. Labor reductions in one area are often offset elsewhere. While banks no longer need large numbers of tellers in the wake of the proliferation of ATM machines and online banking, some have used the savings to open additional convenient neighborhood branches and have added new skills and responsibilities for tellers. The U.S. Bureau of Labor Statistics says that U.S. employed 378,000 bank tellers in 2021, but projects only a 12% decline by 2032.
Gordon (p. 599) points out that search engines, basic AI, and big data analytics have been around for decades and that the quantity of data processed has risen exponentially. It has been long enough, he believes, to expect to see a measurable effect in the productivity numbers, but the case has not been made.
In sum, Gordon is not denying the impact that the ICT revolution and emerging new technologies have had, and will continue to have, on the quality of our lives. He is saying, instead, that this has occurred without countering the sluggish growth in the broader economy or the effects of the various headwinds. Market economies are based on profit, not quality of life. AI, robotics, and big data analytics will only be extensively deployed in the private economy generally where there are profits to be made and, so far, this has largely been confined to areas like e-commerce, advertising, investment banking, and revenue management. Improved living standards for the general population are not built on a foundation of Facebook advertisements.
Gordon paints a bleak global picture, seeing little in the way of innovation on the horizon that might spur a significant rise in productivity and mitigate the headwinds He predicts (pp. 637-639) that over the 25 years from 2015 to 2040, annual real growth rates in all relevant categories in the U.S. and elsewhere will decline significantly in comparison to earlier periods. Stagnation will be with us for a long-time.
In generational terms, this means that the children of baby boomers, i.e., early Millennials and Generation X, and their children, will be facing the brunt of the headwinds with limited resources. They will be supporting, in one way or another, the huge but increasingly unproductive older boomer generation even as they fall short in saving for their own retirement. In the U.S., we are already seeing the result: young adults living with their parents, marrying late in life or not at all, having no or fewer children, avoiding debt, delaying technology upgrades, and hesitating on big purchases, especially homes and new cars. Numerous reports confirm that half of pre-retirees feel they do not have sufficient savings to support retirement, the very definition of economic insecurity. Europe has social security systems that mitigate many of these concerns, but they are fraying rather than improving as workers protest Macron’s retirement reforms in France and younger workers relegated to the gig economy endure lower wages and limited benefits.
Some predict that the children and grandchildren of boomers will be the first generations to have a lower standard of living than that of their parents and grandparents.
Implications for social democratic policy
Gordon’s thesis is important because he is saying that the stagnation problem is not just lagging demand but includes serious barriers to creating new demand which, in turn, causes a decline in potential private sector investment. Gordon’s own prescriptions are not particularly radical, though (pp. 643-645). They involve policies that encourage investment, like capital gains tax cuts and relief from regressive regulations, and policies that help prepare a work force that can implement innovation in coming years, such as more pre-K education and free public higher education. He also wants to increase R&D funding generally and maximize consumer throughput through an increased minimum wage and enhanced low-income cash benefits.
Stimulus is important and center-left politicians and economists across the advanced nations are promoting large-scale public spending. Joe Biden’s inflation Reduction Act will provide hundreds of billions toward infrastructure repair and modernization and will join Europe with heavy investment in green energy and low-carbon industry.
We must address climate change, of course, and improve our bridges, tunnels, and roads, but, based on Gordon’s analysis, this will not directly reverse social-democratic political decline because the spending does not address the root cause of secular stagnation within the framework of poorly-performing capitalist market economies.
With no substantive solution to stagnation on the horizon, simply asking workers to have faith that such spending will address their immediate perceived needs and improve their living standards, is wishful thinking. We can only re-establish a social-democratic consensus by investing directly in people, in housing, livable communities, convenient transport, health care, education, and social security. Failing that, workers most affected by stagnation and rising economic insecurity will continue migrating to reactionary populist parties.
The next post will elaborate further on the perceived needs of working and middle-class families and provide a social-democratic definition of “living standards”.