Sticky Wages as Employment Gaps Persist

Summary

• A new NBER working paper compares today’s industry and occupational composition of the US labour market to pre-pandemic.
• It finds that the pandemic has increased labour market frictions, making it harder for firms to match with workers and vice versa.
• The reason is likely to be a combination of workers trying to upskill and firms looking to fill gaps created by experienced employees leaving.
• Promisingly, most of the temporary labour market issues created by the pandemic have subsided, and the overall industry composition of jobs is unchanged.

Introduction

Wages have been growing uncharacteristically quickly in recent months. This is mainly a by-product of extreme tightness in the labour market, which is most evident in demand-side slack measures. Moreover, current estimates of the natural rate of unemployment suggest it has risen from 3.6% to 4.9% over the last couple of years. The latest U3 print is 3.7%, indicating an overheating economy.

It takes no expert to find the main culprit. The labour market movements early in the pandemic were stunning in both their magnitude and swiftness. And, two years out, the employment-to-population ratio remains markedly below pre-pandemic levels. At the start of 2022, around 7.3mn workers were missing relative to February 2020. Where have they gone? A new NBER working paper investigates.

Pandemic-Induced Temporary Layoffs

The pandemic was characterised by an increase in temporary layoffs. Normally, around a third of layoffs would seek alternative work. But only 10% did in April 2020, meaning employers and workers wanted to maintain their connections despite needing to separate at the start of the pandemic.

CPS longitudinal data shows this. Those who reported being temporarily laid off and were not searching for work in April 2020 (‘Waiting’) had a higher ‘industry return rate’ than those classified as unemployed and searching for work (‘Searching’) in May or June (87% vs 68%).

This ‘Waiting’ fraction, which made up 8% of the adult population, decreased dramatically over the next six months. By late 2020, nearly all the temporarily laid off workers had returned to employment (Chart 1).

A Rise in Labour Force Exiters

One group that remains elevated, even until May 2022, is ‘Not In Labour Force’ (NILF). These are not waiting, searching, nor in employment. The rise in this group coincides with a decline in the employment-to-population ratio (epop), which is still down 0.7 to 1pp from pre-pandemic levels (Chart 2).

What accounts for the missing workers? Fixing the population structure as of February 2020 (i.e., the counterfactual had the age structure remained the same), the epop ratio would be 60% closer to pre-pandemic levels (Chart 2). That is, the aging population can account for a substantial portion of the shortfall in 2022.

This shortfall could have been predicted, but not the other 40%. This group, the authors find, has come from excess declines in employment among the over 65s. There were roughly 1.3mn excess early retirements between February 2020 and December 2021 against the normal number of retirees.

Combined, population ageing and early retirements can explain most of the labour force decline up until now. Other concerns, such as Covid-19, immigration, reduced wage incentives and vaccine mandates, have now broadly faded away.

Revisiting Labour Market Slackness

The previous two sections showed the labour force decline during the pandemic was largely made up of those temporarily laid off and those opting to leave the labour force. Both groups, by definition, do not actively search for jobs. As a result, the number of jobs per job seeker, a key measure of labour market tightness, increased dramatically.

The key, though, according to the authors, is recognising that this uncharacteristically tight labour market is because of the rise in temporary layoffs. Take this group away, and the ratio of unemployed job seekers to job vacancies – the Beveridge Curve relationship – is not too different from pre-pandemic (Chart 3).  

It suggests that the increase in wages is more to do with a decrease in matching efficiency (equivalent to an outward shift in the Beveridge Curve) than the drop in job seekers. This implies greater labour market frictions, e.g., skills mismatches, making it harder for a firm to match with a worker and vice versa.

Does this mean that the types of jobs have changed during the pandemic? That is, are the job posts on offer asking for different skills compared with before the pandemic? Or are workers trying to upskill within the same industries, meaning firms are left choosing from lower skilled workers?

A Reallocation of Workers

Given the changes in workplace practices and acute impact on consumer demand in particular sectors, many believed the pandemic would result in a large reallocation of workers across industries. According to the paper, this has not been the case.

The authors look at the level of reallocation across both industries and sectors (Chart 4). At the start of the pandemic, the composition of jobs changed dramatically as roughly 20mn jobs were lost in particular sectors. However, the immediate reversion to levels seen previously means the structure of employment has not changed dramatically.

The authors also find that the industry return rates – the fraction of hires that return to their previous industry – was higher across all unemployed groups relative to pre-pandemic levels. That means there was less reallocation of workers across industries than usual. In other words, workers were tending to return to the same industry more than in normal times. 

One thing that has changed, however, is the proportion of workers in low-skilled occupations (Chart 5). Despite vacancies for these types of jobs being above pre-pandemic levels, the employment share never recovered. Meanwhile, the share of professional jobs has increased.

So, while industry composition has not changed significantly, the skill composition has. Recognising there are some caveats, this loosely implies that workers are seeking out better skilled jobs in similar industries. This may also explain why firms are generally recruiting into roles with lower education and experience (Chart 6).

To test this, the authors look at CPS data from 2015-2022. It allows them to track monthly transitions in employment, specifically whether individuals are moving between occupations. The results show that blue collar, low-skill service workers and the unemployed were all 10% more likely to move into a professional occupation during the pandemic relative to pre-Covid.

Bottom Line

The results from the paper suggest any transitory pandemic-induced changes to the labour market appear to have subsided. But importantly, the lower employment-to-population ratio is likely to remain. We can reasonably assume, then, the natural rate of unemployment has risen and that a tight labour market will remain in the near term.

Bigger picture, there appears to have been no significant reallocation across industries, but perhaps an upskilling of the workforce. This makes sense as firms look to fill the relatively large gap created by early retirements. As for first-order wage inflation, the paper appears to suggest this should taper off from now. However, if price inflation remains elevated, we could see second-order wage inflation kick-in.