How Accurate Is the Missing Workers Narrative?

There has been a steady drumbeat of comment around labor shortages since economies reopened following COVID-driven lockdowns. Ubiquitous “help wanted” signs; slow service at restaurants; months-long waits for basic installation services; interminable lines in seemingly every checkout lane; flights cancelled due to lack of crew. Is it any wonder that many are asking the question: Where have all the workers gone?

But we must also ask how real is the missing worker phenomenon? Are our frustrations causing us to extrapolate too much from limited individual experiences? After all, a bad customer service interaction is remembered far longer than a good one. When we examine the actual data, we see that the reality differs from the perception — as it often does. Some workers have yet to return to the labor force, the oft-referenced retirement wave is real, disability incidence has risen, and migration flows are far from normalized. There has also been a shift of labor away from some sectors to other areas of the economy. But, by and large, the missing worker phenomenon is not nearly as prevalent as one might think. And most importantly, labor supply has continued to improve and should gradually alleviate the acute wage pressures observed over the past year.

1. The Workers Are Actually Back at Work

Perhaps the most powerful counterargument to the missing workers narrative is that across large developed economies employment is actually back to pre-COVID levels. Taking December 2019 as the benchmark, employment is up 4.6% in Australia, 2.5% in Canada, and is incrementally higher in the United States and Germany. Even among the laggards (Japan and UK) the shortfall is less than 1.0% and shrinking (see Figure 1).

2. Trend Comparisons Are Misleading

There has been an abundance of commentary around how current employment levels remain far below pre-COVID trend. This is essentially an argument of what could have been, but it is misleading insofar as employment occurs in the context of a certain level of economic activity, not irrespective of it (assuming no government interference). If real GDP is not back to trend, why should we expect employment to be? Moreover, it is not just about the number of people working — rather it is about the actual work effort, as measured by total hours worked. This is an important difference. In the US, for instance, employment is only about 0.5% above the December 2019 level but aggregate hours worked are up 1.8%. What explains the difference?

3. A Shift Toward Full-Time Employment

Relatively speaking, the pandemic has triggered a broad shift away from part-time and toward full-time employment, a trend evident across many developed economies (see Figure 2). There are several intuitive explanations for this. At the core is the idea that the economic benefits of a job would need to be substantial enough to warrant the health risks associated with the pandemic, especially in the context of augmented unemployment benefits. This is more likely to be true in a full-time job than in a part-time job. To the extent that medical benefits are linked to the number of hours worked, this may have also heightened the preference for full-time employment. Meanwhile, childcare and other responsibilities may have interfered to a greater degree with part-time workers’ ability to work, since these responsibilities likely shaped their preference for part-time employment in the first place. Given these responsibilities, some households may have chosen to consolidate employment, i.e., move from two individuals with part-time jobs to a single person working full time. Whatever the reason, this distributional shift and increase in average hours worked by each employed person further diminishes the underlying extent of the perceived worker deficit.

The part-time employment deficit is the greatest in the US, and an important related question is whether this decline was voluntary. There are important gender nuances here. Women account for the overwhelming majority of the decline in US part-time employment during the pandemic, perhaps not that surprising given that women made up some 63% of this group pre-COVID (see Figure 3). As female full-time employment has since made new highs, it is probably true for some women that this shift was voluntary, in response to better opportunities.

However, there are other women for whom part-time employment is the only feasible option due to other responsibilities (childcare, etc.). Indeed, when juxtaposed with labor force participation data that show women in the 20–34 age group have lagged their younger counterparts in rejoining the labor force (see Figure 4), one could infer that at least some of the decline in part-time female employment has not been voluntary. This may reflect a story of two labor market realities. For women unencumbered by other obligations, solid employment opportunities abound; but for those subject to such constraints, even the opportunities of the past may now be hard to take up. A very interesting observation here is that men aged 25–34 have lagged all other age and gender cohorts in returning to work. We can’t help but wonder whether this group of men — more likely to be parents of young children — are similarly burdened by family responsibilities. Perhaps the intensity of COVID disruptions to childcare and school schedules have required them to now share more fully in these parenting tasks.

4. Retirement Wave

The data in Figure 4 show that US women aged 45 and over have fared even worse than the 20–34 age cohort in terms of returning to the labor force; women aged 45–54 particularly so. This is not true of men in this age group. In the aggregate, the data is consistent with the idea of a retirement wave, but this seems to be primarily a women’s story. This likely results from women being disproportionately employed in the service industries most acutely affected at the height of the pandemic (leisure and hospitality, in particular), while men are overrepresented in areas such as construction and trucking, where labor demand has been extraordinarily high. Moreover, there was a drastic change in working conditions in areas like education and healthcare (which have a high proportion of female occupations) through this period, which possibly further contributed to early retirements. It remains to be seen how much this early retirement trend can be reversed in coming quarters.

Data from other countries also offer some support for the idea of a retirement wave. However, the trend is not universal, with Australian data hinting at the exact opposite. Moreover, while both US and Canadian data show a decline in the labor force participation of those 55 and over since the onset of the pandemic (all data in Figure 5 is indexed to December 2019), it seems that this is accounted for by those 65 and over. Lack of identical statistics limits a detailed apples-to-apples comparison across countries, but the available evidence points to a modest retirement wave largely driven by those 65 and over. It is not so much that more people are retiring, but perhaps that fewer people of retirement age are choosing to remain in the labor force. Strong financial market returns since the early days of the pandemic may have facilitated these decisions by boosting retirement nest eggs and reducing the need to continue working past retirement age. This too may change, depending on future market conditions.

5. A Lot More Churn

One defining characteristic of the post-COVID US labor market has been the abundance of employment opportunities. With the exception of Japan, job openings are far above pre-pandemic levels. In Australia, they are nearly double what they were in December 2019, while in the US and the UK they are more than 50% higher. This has led to greater worker willingness to quit jobs for better opportunities elsewhere. In the United States, for instance, while quits peaked in November 2021, they remain roughly 700,000 above 2019 averages. For employers, these departures are painful because they are highly unpredictable and require immediate adjustments to ensure business continuity. We suspect that the cumulative impact of higher quits explains a big chunk of the spike in job openings since 2021.

6. Disability in the Workforce

Inability to work due to health problems and long-term disability has also restrained employment. In the United States, for instance, there are about two million more people with a disability than recorded just before COVID, but the number of persons with disability who are employed has increased by only about one million.

7. Sector Rotation

COVID caused a major demand shift away from services towards goods consumption. At its peak in the second quarter of 2021, real goods consumption in the US was 18% higher than at the end of 2019 and remained 14% higher by the second quarter of 2022. A lot of resources, including labor, were required to meet this surge in demand. Unsurprisingly, employment soared in occupations related to warehousing, trade, and transportation, with some sub-components within that category up over 20% relative to end-2019. By the time other service industries like leisure and hospitality were ready to again hire in earnest, they found many of their former employees already gainfully employed elsewhere.

8. Migration Decline

According to the International Labor Organization (ILO), there were 169 million migrant workers in 2019, accounting for 4.9% of the global labor force. Almost 25% of them were in Europe and more than 22% in North America. The pandemic has massively upended global migration patterns. The initial common impact was a sharp decline in net migrant inflows, although the speed and magnitude of healing has since diverged greatly across countries. For example, net international migration to the US was only 247,000 between 2020 and 2021, almost half of the level between 2019 and 2020. Canada also experienced a dramatic decrease during the early phase of COVID, but flows have since rebounded sharply. During the pandemic, Australia experienced its first negative net overseas migration since just after World War II — however, a visible process of normalization is now also under way there. This gives good reason to anticipate that, in time, migrant flows will return to more typical patterns. It is also important to recognize that COVID was not the only shock to have caused major changes to migration patterns. War and the associated influx of refugees into Germany in 2015 is another recent example. Regulatory changes can also play an important role. All this is a reminder that extrapolating from extreme events may prove to be erroneous down the line.

9. Lingering Disincentives to Work?

Some readers may find this particular point insensitive or politically incorrect, but it seems persuasive to us when considered through the objective lens of economics and the idea of rational, utility-maximizing individuals. For a sub-group of potential workers, practical financial considerations surrounding what is generally thought of as an individual fiscal cliff may have played a role in the decision to not rejoin the labor market. An individual fiscal cliff describes situations when the income associated with a new job disqualifies an individual from receiving government financial assistance or diminishes the value thereof enough that employment becomes unappealing. This was the case with many workers in lower-paid occupations during the period of time when augmented unemployment benefits were made available. While those have now ended, other benefit programs remain that may have a similar, if less acute, effect. Data from the US Department of Agriculture (USDA) show there were 41.2 million individual recipients in the SNAP (Supplemental Nutrition Assistance Program) program as of April 2022, 4.0 million more than in December 2019. There were 21.6 million households, which is 2.6 million more than in December 2019. The average household monthly benefit size has also roughly doubled during the same period. It seems difficult to believe that this combination would have had no impact whatsoever on individual decisions around employment among recipient groups.

Labor Market Frictions to Subside

While a degree of worker deficit does indeed exist, its magnitude is not nearly as substantial as generally believed. No single reason explains the phenomenon, with numerous factors playing varying roles over time and across geographies. A substantially higher degree of churn in the labor market tends to exaggerate perceptions around the intensity of this phenomenon, but may now be starting to subside. We anticipate further reductions in labor market frictions from here on, bringing down wage inflation, allowing unit labor costs to decline, and aiding in the process of returning inflation closer to pre-pandemic trends.