|
|
|
|
|
|
|
|
Demography AI and the Future of Economic Growth
Technology Briefing |
|
TranscriptAs discussed in our first four trends this month, the world is now gripped by a once-a-century discontinuous transformation; that transformation conforms to a reliable demographic model which repeats every four generations. This pattern was first documented by historians Neil Howe and William Strauss in their landmark 1990 book, Generations. They referred to these rare and extreme realignments as "fourth turnings" and they result in the birth of a new consensus well-suited to the next era. Such traumatic realignments were associated with the Revolutionary War, the Civil War and the Great Depression era. Demography documents the characteristics of a human population. These include the number of people, as well as their locations, ages, genders, educations, races, and other characteristics. And as documented in prior issues, "demography is destiny," at least in the long run. That's because it drives both the demand and supply sides of the economy. Combined with available natural resources, technologies and the institutional systems which influence human behavior, demography determines what, when, where, how and why we produce and consume everything that makes up our economy. Generations inevitably vary in size and "life experiences." And each generation possesses both cohort and life-cycle characteristics. Cohort characteristics define the members of a generation as being different from the members of the preceding and subsequent generations; they also define the factors that make people similar to others in their generation. For instance, most Baby Boomers share cohort characteristics which differ from those of members of the Silent Generation or Gen X. These characteristics don't tend to change. On the other hand, life-cycle characteristics change in predictable ways as a generation progresses from birth-to-death. The members of a generation can be expected to produce, consume and participate in society in certain predictable ways at various points in their lives. As also documented in Trends and in our book Ride the Wave, technology which enables humans to produce, consume and "participate in society," advances in well-defined waves, referred to as techno-economic revolutions. The long-term progress of our civilization and economy is determined by the way generational patterns in demography interact with advancing waves of technology. The Limits to Growth and the broader "green movement," assume that population growth and rising affluence will inevitably outpace technological innovation. Fortunately, over the past 250-plus years, history has consistently proven this hypothesis wrong! Managers, investors and consumers need to recognize that their careers and lives are ultimately shaped by these demographic and techno-economic trends. However, their fortunes will only be buffeted by the short-term business cycle although that's what most pundits and consultants tend to emphasize to the exclusion of all else when formulating prognoses. That's why it's important to step back and consider how interactions between the digital techno-economic revolution and prevailing demographic trends are creating opportunities and threats even as we're all buffeted by the ups and downs of the business cycle and the new American consensus takes shape. We need to recognize that the emergence of a new American consensus for the 21st century as discussed in trends #2, #3 and #4 this month are necessary to transform society so it can make the most of the opportunities created by the Fifth (or Digital) Techno-Economic Revolution. As we've explained in previous issues, this future looks extremely bright for those who are ready to seize the emerging opportunities and avoid the pitfalls. Consider the facts. Charts in the printable issue clearly show that growth in working age populations around the world has declined precipitously since 2000. However, unlike most of the world's major economies, the United States will continue to grow its working age population through at least 2030. In the United States and elsewhere around the world we're seeing a sharp rise in the "age dependency ratio" caused by retired workers increasing at a far faster rate than the working age population. After reaching a minimum around 2010, this ratio is projected to explode as Baby Boomers continue moving into retirement. Making things worse, labor force growth has not kept up with the growth of the working age population for various reasons discussed in prior issues. Part of this is the generous welfare state. Other factors include a skills mismatch, the pernicious impact of substance abuse and crime, and a pervasive sense of global "disappointment" among millennials. These factors have combined to create an unprecedented labor shortage that began in 2017 and is only likely to get worse. To put it simply, there are more job openings than there are people to fill them. As of year-end 2021, U.S. employers had 10.9 million vacant positions. Even though that was nominally below the record 11.1 million from earlier in 2021, there was an unprecedented 1.7 open jobs for every unemployed worker. Data shows that America's long-term growth rate in per capita GDP has slowed along with the decelerating growth of the working age population. What is this telling us? The world is going to experience a rising dependency ratio for the balance of the 21st century. That means fewer workers producing to meet the needs of a larger dependent population. Conventional thinking assumes that this will inhibit the expansion of both output and consumption, limiting economic growth. That's because economic output is inherently tied to population while older people tend to consume less. But fortunately, the fifth techno-economic revolution seems poised to challenge that conventional wisdom. On the surface, America's barely growing work force appears to create a major barrier to strong economic growth in the 2020s and beyond. However, the unique characteristics of Artificial Intelligence are enabling capital to replace or enhance human functionality in ways previously impossible. That will mean lower unit costs enabling increased global demand, while potentially providing better, more rewarding jobs. According to research from ARK Investment Management, the value of a knowledge worker empowered with AI will increase at a 15% annual rate during the next eight years. According to their assessment of broad categories of knowledge workers, AI software could cumulatively boost the productivity of the world's average knowledge worker by nearly 140%, adding approximately $50,000 per year in value per worker. Based on 1.1 billion knowledge workers worldwide, that's a total of $56 trillion in additional annual value creation, globally. By comparison, consensus expectations assume 2.7% annualized real wage growth through 2030. Anticipating more than a doubling in productivity, ARK forecasts that companies will be willing to spend roughly 25% of that gain on AI software. In other words, for every $1 gain in productivity, they should be willing to dedicate $0.25 to software licensing, most likely in the form of Software-as-a-Service. ARK estimated the value of the productivity gains based on the "automatability" - and productivity multipliers of various labor tasks relative to the wages associated with several job categories, discounted to account for lower international wages. Research by McKinsey & Company (in 2017) also analyzed the proportion of labor tasks AI could automate by 2030. And they formulated a "mid-point scenario" which suggests a level of about 15%. However, ARK's updated assessment indicates that the displacement could actually be much higher for two reasons: 1) technological advancements and 2) the adoption curve. And by 2030, this could reach 60%- plus for roughly half of U. S. knowledge workers. The current slope of AI advancement is very steep because innovations in hardware, training methods, and neural network architecture are compounding to accelerate progress beyond what Moore's Law would predict. For instance, OpenAI, DeepMind, and other organizations have demonstrated that AI models should be able to achieve near human-level proficiency in many narrow knowledge worker tasks. ARK's research also suggests that AI training costs are dropping at a rate of roughly 60% per year, potentially breaking down barriers to many exciting "large model" projects. And it looks like, higher allocations of human and financial capital to AI projects will continue to accelerate this rate of innovation. Accelerating the adoption of AI software offers a clear return on investment, particularly because of the acute shortage of specialized knowledge workers. For example, an AI tool that doubles the productivity of engineers is a compelling value proposition for software companies struggling with recruitment. And failure to adopt this technology is likely to put many companies in harm's way. Furthermore, the impact of the AI-based productivity revolution won't be limited just to offices and knowledge workers. Factories and farms are going to feel its impact too. And that's one reason that reshoring manufacturing and logistics makes so much sense. Companies across North America laid out more than $2 billion for 39,708 robots in 2021, 28% more than in 2020. That's an all-time record and the numbers in 2022 are likely to be even stronger. These were intended to help them contend with record demand and the pandemic-fueled labor shortage. Robots went to work in a growing number of industries, expanding well beyond their historic role in the automotive sector. In the farming sector, John Deere recently began offering an upgrade to make its top-of-the-line tractors, fully autonomous when tilling fields. Meanwhile, venture capital is rapidly flowing into start-ups related to creating a wide range of harvesting machines for fruits and vegetables. Vision systems, chemical sensors and tactile sensors are combining to open up a wide range of applications to automation. As with office work and manufacturing, a quantum leap in farm productivity is at hand. What's the bottom line? A demographically driven labor shortage is hitting every major economy. But fortunately, a revolution in the cost and performance of AI-based technology is likely to more than offset this shortage. The challenge is for companies to embrace this technology in ways that maximize profits, while minimizing disruptions for workers and consumers. Given this trend, we offer the following forecasts for your consideration. First, demography will provide a reliable context for economic decision-making. It enables us to reliably forecast the number of workers and consumers by age, location and gender for the next 20 years. Obviously, many factors will influence their behaviors. However, we know that it's unlikely that statistically significant numbers of people will suddenly appear or disappear. Second, unlike most countries, the United States will have the option of accessing an enormous supply of migrants eager to join our workforce. However, the tradeoffs are complex. For the moment, we need to assume that legal immigration will remain capped and illegal immigration will be curtailed and even reversed. Third, over the next two decades, the growth-enhancing power of technology will overwhelm the growth-killing power of demography. Unlike the general-purpose technologies underpinning previous techno-economic revolutions, AI and robotics will be able to substitute for human labor in previously unimagined ways. The result will be accelerating economic growth despite extremely slow growth in the U.S. workforce. Fourth, solid economic growth will enable us to provide a strong social safety net for older Americans and those who are genuinely disabled. Assuming they aggressively embrace digitization, the OECD countries and the United States in particular, will be able to meet the needs of their aging populations. Because of its scale, natural resources, and entrepreneurial culture, the United States is uniquely well positioned for this new era. One of the biggest challenges will be creating incentives for America's NEETs and China's "lying flat generation" to become contributing members of society. Fifth, even if our stated assumptions regarding automatability, productivity-multipliers and adoption curves prove optimistic, AI and robotics will still prove transformational over the coming decade. |
|
Comments
|
|
|
|
|
|
|
|
|