Sources of Labor Market Upheaval — and How the Workforce Can Adapt to Constant Change
2020 has been a year of unprecedented change, with a global pandemic temporarily putting entire economies on hold and the death of George Floyd painfully reminding us of the deep-rooted racial injustice and inequality in this country and around the world. For many in the United States, the pandemic has meant either having to adapt to working from home, for the lucky ones, or being furloughed or laid off.
COVID-19 has affected people extremely unequally: Individuals with less than a high school degree experienced unemployment rates three times as high as individuals with a college degree (BLS, 2020). Economically, the virus has particularly hurt workers in occupations requiring close human interactions such as personal care or food services (BLS, 2020). There is evidence that people with high incomes and good internet connections have been better able to stay home during this crisis. While these may be circumstances specific to a global pandemic, the unequal effects are not new. Research shows that workers in occupations with low and middle education levels have been repeatedly adversely affected by labor market shocks over the past 30 years.
Earnings and wage inequality have been steadily rising in the United States since the 1980s, with stagnating wages at the bottom end of the wage distribution being a major factor (see, e.g., Acemoglu and Autor, 2011, for an overview). While household incomes in the top one-third of the distribution have grown by 8% between 2000 and 2018 (and receive 48% of aggregate incomes), the bottom one-third has seen household incomes grow by only 1.7% (and receive 9% of aggregate income) (Pew Research Center, 2020). This gap becomes much larger when one considers smaller slices further apart in the distribution: Hourly wages at the 95th percentile of the wage distribution have grown by 25.1%, whereas those at the 10th percentile have grown by only 6.9% in the same timespan (Gould, 2019). It is also worth noting that the overall labor share — the share of GDP coming from labor income — has been continuously falling.
There are many contributing factors and partial explanations for this increase in inequality. Sources of changes to the labor market — factors such as changes to the number of jobs, which occupations are in demand, and how work gets done — contribute extensively to the problem. Major sources of labor market change are usually studied in isolation from one another: technological progress, trade with China, plant closures, recessions, the global pandemic, and other factors. Labor economists usually study these phenomena and their effects separately, attempting to isolate each cause. However, it is also helpful to step back and compare the findings across many studies, in order to look at all the sources of change together: who they affect, in which way, and how their effects interact, both with one another and with other factors in the economy.
We can glean two broad sets of facts from this exercise:
- We learn which groups have been most affected by different shocks to the U.S. labor market. Often, the same or similar groups — in terms of occupations, education levels, race, demography, and geography — have been repeatedly affected adversely by shocks, possibly explaining why they may be struggling more than others. These groups are particularly vulnerable today and need appropriate economic policies to support them.
- We can also learn which groups have been able to overcome adverse shocks and how they have been able to adapt. We need to emulate the experience of these individuals and learn from them in order to derive the right policies for the first group.
What follows is a summary of the empirical evidence about the major sources of labor market change in the United States over the past 30 years. If 2020 has shown one thing, it is that the only constant in our current economy is constant change, and the workforce needs to be able to adapt. However small or large the next change or shock may be, we need appropriate policies that both enable adaptation and help those who struggle to adjust on their own.
Technological progress
Technology and automation have decreased demand for routine-intensive jobs and increased demand for tasks that are complementary to technology.
Technology is a major factor in disrupting the labor market because it changes how work gets done. The way labor economists typically study the impact of technology on occupations is by determining what share of tasks in a particular occupation lends itself to automation. For example, technology enables the automation of both physical and mental routine tasks, which were formerly performed by lower- to middle-skill workers, therefore decreasing demand for such workers (e.g. Autor and Dorn, 2013).
Contrary to the effects of trade (discussed in the next section), technology does not have strong overall employment effects on local labor markets (Autor, Dorn, and Hanson, 2015). Automation instead polarizes the labor market: Employment in routine-intensive occupations, which are usually middle-skill occupations, declines, while employment in both high-skill and low-skill non-routine-intensive occupations increases (Autor, Levy, and Murnane, 2003; Goos and Manning, 2007; Autor and Dorn, 2013; Michaels, Natraj, and Van Reenen, 2014). Non-routine-intensive high-skill occupations are typically characterized by abstract and analytical tasks, while non-routine-intensive low-skill occupations are manual occupations that cannot be easily automated, such as cleaning or repairs.
Of course, “technology” is a catch-all phrase for all sorts of innovation. What technology can do has been changing — and with it, the tasks it may take over and occupations it can affect. For example, while it was previously thought that only highly repetitive tasks in a fixed environment lent themselves to automation (an idea that much of the older labor economics literature on automation relies on), newer AI-augmented technologies allow the technology itself to learn with more data. This means the technologies are increasingly able to deal with complex tasks in changing environments.
A recent study suggests a framework for analyzing the effect on the labor market of different technologies — such as robots, software, and AI — by comparing patent texts of those technologies with occupation descriptions and calculating the extent of potential exposure of an occupation to the technology in question (Webb, 2019). Robots are good at very different tasks than AI: Robots can repeat manual tasks or carry heavy objects around, while AI enables computers to analyze, predict, and recognize patterns — all while getting better at it over time with an increasing amount of data. It is therefore unsurprising that the study finds the most exposed occupations to robots — miners, janitors, housekeepers, maids, butlers, and cleaners — to be manual and typically low-education occupations requiring physical strength. Large shares of tasks in these occupations could be replaced by robots. On the other hand, the most exposed occupations to AI include financial advisors, optometrists, physicists, and astronomers — all high-education occupations in which AI’s distinct capabilities at analysis and decision making could partially replace tasks currently performed by humans. Some researchers therefore believe that AI technology has the potential to actually decrease inequalities by levelling the playing field across occupations. Even if one could not entirely replace humans with technology in those occupations, the total demand for humans in the exposed occupations could change drastically as a result.
Since different technologies affect different tasks and occupations, their impact on wages and employment likely differs. One study looking at the adoption of industrial robots and its impact on local labor markets in the United States (Acemoglu and Restrepo, 2020) shows that adding one robot replaces 3.3 workers on average. Because of the nature of work being replaced, the burden falls on low- and middle-skill workers. The overall employment effects of AI have yet to be analyzed in greater detail, but one study suggests that many occupations are only partially replaceable by AI, which may give rise to reorganization and rebundling of tasks into new occupations (Brynjolfsson, Mitchell, and Rock, 2018).
Most occupations are affected by technological progress in some way, just as our daily lives are. Everyone from supermarket cashiers to doctors needs to frequently adapt to new tools in their jobs. One paper looking at online job postings in STEM occupations shows that technological change continuously introduces new required skills and makes old ones obsolete. This explains the high initial but then flatter returns to STEM majors in the labor market (Deming and Noray, 2020). The main takeaway from studying the impact of technologies on labor markets is that it is beneficial to have complementary skills to whatever technology is good at, in order for one’s work to be augmented, rather than replaced, by technology. Unsurprisingly, therefore, social skills have especially been rising in importance in the labor market (Deming, 2017).
Trade with China
The impact of globalization on local U.S. labor markets, in particular with the rise of China — what the labor economics literature calls “the China trade shock” — is a well-studied phenomenon. Trade shocks from China in particular hurt manufacturing workers in regions specializing in labor-intensive industries, concentrated in the U.S. heartland.
Imports from China as a percentage of U.S. economic output doubled within four years of China joining the World Trade Organization in 2001. In their seminal study, Autor, Dorn, and Hanson (2013) find that increased import competition from China can explain up to 40% of the total decrease in manufacturing employment in local labor markets and of the corresponding growth in local unemployment and non-employment between 1990 and 2007.
Figure 1 shows which manufacturing industries saw the largest increases in import competition from China: electrical equipment and textiles industries, both labor-intensive industries that manufacture goods in which China became competitive. The U.S. regions that were the most impacted by the China trade shock were therefore regions specializing in these strongly affected industries. They are concentrated in South Central states, including Tennessee, Missouri, Arkansas, Mississippi, Alabama, Georgia, North Carolina, and Indiana. The labor markets most strongly affected were characterized by below-average education levels and higher unemployment levels even before the shock, which is why these regions were so detrimentally hit and have been struggling to recover ever since.
How has this affected individual workers? Workers who were employed in import-intensive manufacturing industries before 1991 experienced lower cumulative earnings between 1992 and 2007, a higher probability of receiving disability benefits, and a higher probability of switching employers, both within and outside of manufacturing (Autor et al., 2014). The effect is more pronounced for individuals who had low initial wages, low tenure, and low labor market attachment. While impacted high-wage workers managed to make up lost earnings by moving into other industries, low-wage earners did not manage to offset their earning losses, either within their original industry or outside it. In fact, outside their initial industry, they accrued even larger losses on average. This shows that these workers have skills that are highly firm- and industry-specific, which are hard to apply in other jobs or industries.
The good news is that trade with China has not decreased overall employment in the United States, because the losses in manufacturing employment were made up by new jobs in other sectors. A recent study by Bloom et al. (2019) suggests that large multinational firms offshored their production but created new service-sector jobs. However, these new service jobs were moved out of the U.S. heartland, where the manufacturing jobs used to be, into high-education areas along the coasts, and were therefore taken up by very different people. As the papers on trade shocks show, geographic mobility has not been a major way of adjusting to the shock (Autor, Dorn, and Hanson, 2013; Autor et al., 2014). This means the China trade shock has definitely contributed to the increased geographical disparities in the United States. This raises the question of why workers have not moved to regions with better opportunities in order to adjust. In summary, the impact of trade shows a recurring pattern: Better-off workers are able to recover from adverse shocks by adapting — in this case switching industries — while the low-wage and less-educated workers struggle to adapt.
Displacement through plant closures and mass layoffs
Plant closures and mass layoffs are not a new phenomenon shocking labor markets — they have occurred ever since the Industrial Revolution and are often a result of the other sources of change mentioned previously. It is therefore instructive to look at who recovers from a job displacement following a plant closure or mass layoff and who does not. This can offer valuable insights into what characterizes individuals who can adapt, precisely because there are so many occurrences across time, places, and industries.
Plant closures and mass layoffs have particularly hurt older and long-tenure workers. On average, displaced workers suffer a large earnings shock of up to 60% in their first year of displacement (e.g. Jacobson, LaLonde, and Sullivan, 1993; Couch and Placzek, 2010) and continue to see significant reductions in earnings — around 25% — six years after displacement. These effects have been shown to vary with the business cycle and are significantly larger when there is already high unemployment in the labor market (Davis and von Wachter, 2011). Exactly which group of workers sees the highest losses depends on the context and data used, but one finding is consistent across all studies on plant closures and mass layoffs: Higher-tenure or older workers (or both, as they are often correlated) see the largest losses. What this tells us is that either workers who have been at the same firm for a long time have highly firm-specific skills that are less valuable and applicable outside that particular firm, firms discriminate against hiring older workers, or older workers, in general, have a more difficult time adapting. It is also possible that older workers had been paid above their marginal product in line with an implicit contract where workers earn less when they are younger and more when they are older (to provide incentives to younger workers) with their previous employer, making it hard to now find wages at this higher rate (Lazear, 1979). The resulting takeaway is the same and is the recurring theme of this article: Those who can adapt and transfer their skills to new settings have the best chance to recover from an adverse shock.
Recessions
Recessions are a recurring fact of market economies, and unemployment is a consequence of recessions (so much so that recessions are often measured by the unemployment rate). And they hurt those the most who were already struggling.
As you may expect by now, losing your job in a recession is worse for some groups than for others. While each recession is different, over the past three decades, the most vulnerable populations in recessions have been men, Black and Hispanic workers, and low-education workers (Hoynes, Miller, and Schaller, 2012; Davis and von Wachter, 2011). Recessions can cause long-lasting effects on the lifetime earnings of affected individuals. In the Great Recession starting in 2007, while management, professional, and service occupations saw no great declines on average in terms of employment, the remaining occupation groups (sales and office occupations, natural resources, construction and maintenance occupations, as well as occupations in production, transportation, and material moving) all saw major declines in employment and have not fully recovered. Unemployment rates during the recession also varied widely by education level. While the unemployment rate for individuals with a high school degree was 11% at its peak (in October 2009), it was only 5% for those with a bachelor’s degree or more.
A silver lining of recessions is that college enrollment rates typically increase because people’s outside options in the labor market diminish in value, which decreases the opportunity cost of education. More individuals getting more education, which can shield them from adverse effects of shocks, is generally good news. However, colleges that depend on state funding saw dramatic reductions in state budget allocations after the Great Recession, which accelerated the cost-shifting from public subsidies to individual payments in higher education and increased the burden of student debt (Barr and Turner, 2013).
Global pandemics
Global pandemics are not a natural candidate for the list of labor market disrupters, but they clearly belong on it, as the experience of 2020 has shown. The COVID-19 crisis and the ensuing lockdowns of entire economies undeniably disrupt the labor market — in a way that most affects those who cannot work from home.
In the months following the start of the COVID-19 pandemic, the economy plunged into a deep recession, with economic activity coming to a grinding halt, unemployment at a record 14.7% in April 2020, and millions forced to work from home. We have yet to see how permanent the changes are and how long it will take for the economy to recover — but a few clear trends emerge. First, the crisis has underlined the importance and value of essential workers, especially those contributing to the running of the healthcare system. Essential workers in front-line jobs were the only job categories that did not see a decline in demand, according to online job posting data (Kahn, Lange, and Wiczer, 2020). All other occupations saw a decline in job offer postings of 30% in April 2020 compared to the start of the year — which can be an important forward-looking economic indicator.
Figure 2 shows actual unemployment rates by occupation groups in April 2020, showing that jobs with close client contact such as the service, food, and non-essential retail industries have been hit particularly hard in terms of layoffs. Work that can be done from home has seen fewer adverse impacts from the crisis, for obvious reasons. Education levels also play a major role: The unemployment rate in April 2020 for individuals with at least a bachelor’s degree was half that of those with at least a high school degree and almost one-third of the unemployment rate for individuals without a high school degree.
A new study finds that up to 37% of jobs in the United States could theoretically be done from home (Dingel and Neiman, 2020), meaning that given the occupation tasks, they do not necessarily require a person to be present at the place of work. These tend to be non-manual and high-education “office” jobs. They are also not uniformly distributed across the country: They strongly depend on local labor market characteristics such as industry distribution and average education levels. For example, while more than 45% of jobs in San Francisco, San Jose, and Washington, D.C., could be performed at home, the same is true for only 30% or less of the jobs in Fort Myers, Grand Rapids, and Las Vegas.
In practice, people’s ability to stay at home depends not only on their work, but also on their broadband access. A study of mobile device data finds that individuals in regions with high-speed internet access are less likely to leave the house. They also find that internet access coupled with high income predicts staying at home particularly well—meaning that inequality has been driving people’s ability to self-isolate and stay safe (Chiou and Tucker, 2020). Of course, the ability to work from home has not insulated the workforce completely from layoffs and redundancies — reduced demand for many services due to the pandemic has also meant layoffs of many employees who could have done their work from home.
Findings from a large-scale survey (COVID Impact Survey) asking individuals about their symptoms and work behavior suggests that many individuals go to work even though they are sick — indicating that the CARES Act has not been sufficient to keep sick individuals at home. These findings suggest that the COVID-19 crisis is one more factor that can drive inequality between educational and occupational groups as well as across geographic regions — not only in terms of income and labor market outcomes, but also in terms of health and virus exposure risk.
Furthermore, the pandemic has emphasized — as no other source of change has — the importance of adaptability in dealing with shocks. This shock has forced everyone to adapt — all but essential workers were asked to stay at home. In industries where work from home is a possibility, technology adoption has rapidly accelerated, leap-frogging the use of remote work tools such as videoconferencing and VPN connections by years. Individuals not only had to quickly adopt new technologies, they also had to get accustomed to a new working situation. This entailed a lot of improvising, coordination, and compromising with other family members as workers coped with child-care challenges or simply the distractions of home while trying to work.
The nature of this crisis, with the most affected industries being those in which women are overrepresented, has meant that contrary to “usual” crises, women have lost more jobs than men. It has also been shown that the burden of at-home childcare during the lockdown has fallen disproportionately on women, a new study by Alon et al. (2020) finds. This threatens to widen the gender pay gap and erase years of progress in this domain. However, the authors also see an opportunity for societal change in this crisis: Because women are also overrepresented in essential front-line jobs such as nurses and therefore had to work while their partners were staying home, more men than before have been charged with taking care of children.
Whatever the personal situation and challenges have been to be overcome, as with any change, the companies and individuals that were best able to adapt have been more successful in navigating these difficult times. People working from home are not the only ones who have needed to adapt — some workers switched jobs to work in industries that continued operating during the lockdown and therefore needed to adapt quickly. For example, with demand for transportation and rideshares down and internet orders soaring, rideshare company Lyft partnered with Amazon, referring drivers on its platform to jobs at Amazon as delivery drivers, warehouse workers, and shoppers. While this may be an extreme example, it underlines the importance of adaptability and mental flexibility to succeed in this economy.
Other factors
In addition to the more sudden “shocks” to the labor market described above, there are also phenomena that are steady developments, or that interact with and amplify the factors considered above.
Worker bargaining power
Workers’ bargaining power vis-à-vis their employers has been gradually decreasing. This is a factor in increasing inequality and stagnating wages, especially for the bottom part of the income distribution. In a new paper, Stansbury and Summers (2020) offer empirical evidence for falling worker power and show that it can explain the falling labor share (the part of national income earned through wages) and increased firm market powers. The first aspect of falling worker power is de-unionization: Private-sector union membership has been steadily falling since the 1950s, from 35% to a mere 6% today — and as a result, the bargained union wage premium has also fallen. Secondly, the wage premium of working in large companies, which typically played the same role as being in a union, has also fallen. Lastly, profit sharing between companies and employees has also diminished. The study also shows that rent-sharing ability has decreased more for workers without a college degree than those with one — adding to the divergence between high- and low-educated workers.
Spillover effects in local markets
The region workers live in also affects their outcomes. Local economies are interconnected: Someone can hold the perfect high-paying job that escaped all the shocks mentioned here, yet still be adversely impacted by living in a region that has been strongly affected by the shock. Because of a particular region’s industry composition, it may have been exposed to higher unemployment and changing industry structures. This may affect the livelihood of workers not in the affected industries through spillover effects. Research finds that trade shocks and mass layoffs in local labor markets have substantial spillover effects on employment in firms and industries other than those affected. It has also been shown that individuals can adapt through regional mobility, i.e., moving to less affected regional economies (Gathman, Helm, and Schönberg, 2020).
Labor market regulations
How work is organized is also a product of the laws and regulations of the market in which it exists. Labor market regulations can affect how work gets done, who can do it, and the rights and benefits associated with working. There are many such regulations, and they vary across states. To assess the impact of regulations, these are the questions to consider:
- How easy is it to hire or fire a new worker? The easier it is to hire a worker, the less inclined firms may be to resort to alternative work arrangements such as gig work. The easier it is to fire workers, the faster companies are inclined to let individuals go in difficult times.
- Are there seniority rules to firing? This will change who gets laid off and how they may cope with it.
- How many times can a firm prolong a temporary contract before having to offer a full-time contract? In the same vein, how are gig workers and independent contractors classified? California’s recent Assembly Bill 5 codifies a test to classify workers as employees versus independent contractors and lays the burden of proof on the company when they classify workers as the latter.
- How easy is it to switch occupations or to move to a different state to practice an occupation elsewhere? Licensing laws may pose serious barriers to entry into certain occupations and impede mobility across states as a means of adjustment.
Interaction among sources of change
Sources of labor market change do not exist in a vacuum. They can also interact with, reinforce, or undermine one another. One such example is that mass layoffs affect workers twice as much when they happen in a recession because workers’ other options get diminished. Another example is that individuals enter gig work when they experience major losses from regular wage income (Koustas, 2019), underlining the important interactions of technology, labor laws, and the state of the economy. The impacts of trade and technology can either reinforce or substitute for one another: On the one hand, greater price pressure from import competition may increase pressure to adopt automation to keep prices under control. On the other hand, given that automatable tasks are also easier to offshore (Blinder, 2009) and falling trade costs reduce the cost of offshoring, certain tasks may be offshored rather than automated, depending on the relative costs. In the same vein, research has suggested that robot adoption and the age of the working population are positively correlated across countries (Acemoglu and Restrepo, 2018).
Conclusion: Adaptability for all
All of these sources of change are at least partial culprits for the current state of inequality and unequal economic opportunities among occupations or regions. Technological progress, trade shocks, and recessions have all been driving structural change and de-industrialization, which have affected regions differently depending on their initial industry structures and workforce characteristics.
The worst affected workers for every shock have been the less-educated ones in manual jobs — that is, those who already earn the lowest wages and whose wage growth has been essentially stagnant for a long time. This is the group that would benefit most from well-designed, targeted economic policies focused on enabling adaptability. To mitigate the impact of trade shocks, it has been suggested that existing trade adjustment programs need to be modernized and increased in size, providing wage insurance and additional assistance to displaced workers, in order to more broadly share and redistribute the gains made from increased trade integration (Autor, 2018).
Other examples of possible mitigation policies are offering geographic relocation assistance, decreasing barriers to entry such as occupational licenses to enable occupational mobility, providing better information on which occupations are and will be in demand on a national scale, or offering retraining and upskilling programs. Such training programs may either be taken up by individuals or provided by firms through financial incentives to retrain their existing workers instead of laying them off once their skills become obsolete. For upskilling and reskilling programs, digital and online training programs offer many advantages, including convenience, ease of access regardless of location (conditional on a stable internet connection), cost-effectiveness, and potential for personalization. Digital content development, learning apps, and best practices in online teaching have made major advances in recent years, and the pandemic has further accelerated developments in this domain.
The findings summarized here underline the importance of systematically investigating heterogeneous effects of shocks on the labor market. While the average effect may be manageable and there are individuals who can adapt to the challenge, we need to pay close attention to those who find it the most difficult to adapt. Knowing which subgroup is impacted the most allows for better policies that could support them — or even prevent the strong adverse effects in the first place. Understanding the heterogeneous effects of job displacements systematically, in a data-driven manner using machine learning, is one of the current research projects in our lab.
This review not only sheds light on who suffers the most following labor market changes, but it points to what it may take to overcome adversity in the labor market. It is clear that younger, better-educated workers in high-skill, non-routine, abstract, and analytical-task-focused occupations tend to adapt to change and adversity better than others. This is not a particularly helpful insight in itself, because we cannot simply make everyone younger, make them graduate from college, and put them in management occupations. But we can try to understand what it is that these workers have in common, what the essence of being better able to adapt is. This is an open research question that has not been sufficiently answered yet. New machine learning methods, together with large amounts of individual-level labor market data, allow characterizing and extracting the traits and skills that allow individuals to adapt and transition into new occupations. This is another ongoing research project in our lab.
The evidence so far suggests that the skills enabling individuals to adapt are so-called general skills, as opposed to specific skills. General skills are adaptable and transferable, and they can be applied in many different situations. These skills are complementary to technology and include communication, reading and writing, and people or social skills. These skills make individuals able to move into a new labor market region and thrive there, change jobs or industries, or continuously learn new skills that their current job requires. They have a lot to do with mental flexibility and a learning and growth mindset. If we can understand how to successfully teach these capabilities, starting in high school and continuing in the upskilling and reskilling programs designed to help struggling workers, this would be a major step in the right direction of equipping the entire workforce with the skills needed in a labor market that is constantly changing. The possibilities and opportunities for online training programs and learning apps are making this policy goal more attainable. One thing is clear: The only constant is change, and everyone needs to be able to adapt, from the university professor to the janitor. The entire workforce needs to be equipped with the skills that enable adaptability in the future.
— —
Learn more about the Golub Capital Social Impact Lab at Stanford Graduate School of Business.
Follow us @GSBsiLab.
Written by Lisa K. Simon
References
Acemoglu, D. and Autor, D., 2011. Skills, Tasks and Technologies: Implications for Employment and Earnings, in Handbook of Labor Economics, 4B, pp. 1043–1166.
Acemoglu, D. and Restrepo, P., 2020. Robots and jobs: Evidence from US labor markets. Journal of Political Economy, 128(6), pp. 2188–2244.
Acemoglu, D. and Restrepo, P., 2018. Demographics and automation. NBER Working Paper, №24421 March 2018.
Alon, T., Doepke, M., Olmstead-Rumsey, J., and Tertilt, M., 2020. This Time It’s Different: The Role of Women’s Employment in the Great Lockdown.
Autor, D.H., 2018. Trade and labor markets: Lessons from China’s rise, IZA World of Labor, Institute of Labor Economics (IZA), pp. 431–431, February.
Autor, D.H. and Dorn, D., 2013. The growth of low-skill service jobs and the polarization of the US labor market. American Economic Review, 103(5), pp. 1553–97.
Autor, D.H., Dorn, D., and Hanson, G.H., 2013. The China syndrome: Local labor market effects of import competition in the United States. American Economic Review, 103(6), pp. 2121–68.
Autor, D.H., Dorn, D., and Hanson, G.H., 2015. Untangling trade and technology: Evidence from local labour markets. The Economic Journal, 125(584), pp. 621–646.
Autor, D.H., Dorn, D., Hanson, G.H., and Song, J., 2014. Trade adjustment: Worker-level evidence. The Quarterly Journal of Economics, 129(4), pp. 1799–1860.
Autor, D.H., Levy, F., and Murnane, R.J., 2003. The skill content of recent technological change: An empirical exploration. The Quarterly Journal of Economics, 118(4), pp. 1279–1333.
Barr, A. and Turner, S., 2013. Expanding Enrollments and Contracting State Budgets: The Effect of the Great Recession on Higher Education. The Annals of the American Academy of Political and Social Science, vol. 650, 2013, pp. 168–193.
Blinder, A.S., 2009. How many US jobs might be offshorable? World Economics, 10(2), p. 41.
Bloom, N., Kurmann A., Handley, K., and Luck, P., 2019. The Impact of Chinese Trade on U.S. Employment: The Good, The Bad, and The Debatable, Stanford mimeo.
Brynjolfsson, E., Mitchell, T. and Rock, D. 2018. What Can Machines Learn, and What Does It Mean for Occupations and the Economy? AEA Papers and Proceedings, 108, pp. 43–47.
Bureau of Labor Statistics, 2020, The Employment Situation, May 2020.
Chiou, L. and Tucker, C., 2020. Social Distancing, Internet Access and Inequality, NBER Working Paper, №26982, April 2020.
Couch, K.A. and Placzek, D.W., 2010. Earnings Losses of Displaced Workers Revisited. American Economic Review, 100 (1), pp. 572–89.
COVID Impact Survey, conducted by NORC at the University of Chicago for the Data Foundation.
Davis, S.J. and von Wachter, T., 2011. Recessions and the Costs of Job Loss, Brookings Papers on Economic Activity, Economic Studies Program, The Brookings Institution, vol. 42(2 (Fall)), pp. 1–72.
Deming, D.J., 2017. The Growing Importance of Social Skills in the Labor Market. Quarterly Journal of Economics; 132 (4), pp. 1593–1640.
Deming, D.J. and Noray, K., 2020. Earnings Dynamics, Changing Job Skills, and STEM Careers.
Dingel, J.I. and Neiman, B., 2020. How many jobs can be done at home? NBER Working Paper №26948, National Bureau of Economic Research.
Gathmann, C., Helm, I., and Schönberg, U., 2020. Spillover Effects of Mass Layoffs, Journal of the European Economic Association, Volume 18, Issue 1, pp. 427–468.
Goos, M. and Manning, A., 2007. Lousy and lovely jobs: The rising polarization of work in Britain. Review of Economics and Statistics, 89(1), pp. 118–133.
Gould, E. 2019. State of Working America Wages 2018: Wage Inequality Marches On — and Is Even Threatening Data Reliability. Economic Policy Institute, February 2019.
Hoynes, H., Miller, D.L., and Schaller, J., 2012. Who suffers during recessions? Journal of Economic Perspectives, 26(3), pp. 27–48.
Jacobson, L.S., LaLonde, R.J., and Sullivan, D.G., 1993. Earnings losses of displaced workers. The American Economic Review, pp. 685–709.
Kahn, L.B., Lange, F., and Wiczer, D.G., 2020. Labor Demand in the Time of COVID-19: Evidence from Vacancy Postings and UI Claims, NBER Working Paper №2706, April 2020.
Koustas, D.K., 2019. What Do Big Data Tell Us About Why People Take Gig Economy Jobs? AEA Papers and Proceedings (Vol. 109, pp. 367–71).
Lazear, E. 1979. Why Is There Mandatory Retirement? Journal of Political Economy, 87(6), pp. 1261–1284.
Michaels, G., Natraj, A., and Van Reenen, J., 2014. Has ICT polarized skill demand? Evidence from eleven countries over twenty-five years. Review of Economics and Statistics, 96(1), pp. 60–77.
Pew Research Center, 2020. “Most Americans Say There Is Too Much Economic Inequality in the U.S., but Fewer Than Half Call It a Top Priority.”
Stansbury, A. and Summers L.H., 2020. The Declining Worker Power Hypothesis: An Explanation for the Recent Evolution of the American Economy, NBER Working Paper №27193, May 2020.
Webb, M., 2019. The Impact of Artificial Intelligence on the Labor Market. Available at SSRN 3482150.
