Labor Demand in the Time of COVID-19
University of Rochester, McGill University and Stony Brook University
More than 26 million workers have filed for unemployment in a span of five weeks in the United States. The historic and sudden rise in unemployment across the country is without precedent. Looking beneath the headline numbers can provide a clearer view of the fast evolving situation. Initial Unemployment Insurance (UI) claims are one of the few real-time indicators of the state of the labor market, however they provide only a coarse view. The data reflect job matches that have been dissolved in the recent past and most states do not break down the numbers by industry. Thanks to a novel data source of job vacancies posted on the internet, collected on a daily basis, we have a real time measure of employer demand. These postings data are an inherently forward-looking measure, representing new jobs a firm seeks to fill, which can be disaggregated to the detailed industry-, occupation-, geographic location-, and even firm-level. Together UI claims and job vacancy data provide a more detailed account of how the labor market evolved over the last weeks and how it is likely to continue to evolve.
The labor market collapsed at the same time across the U.S. irrespective of the early spread of the virus or the timing of state-level policies imposed.
- The bulk of coronavirus cases and the harshest policy restrictions did not begin in the United States until the second half of March and were staggered across the states. The epidemic hit the U.S. relatively late. Parts of China were put under quarantine order by the end of January and large parts of East Asia shutdown much of public life in February. In Europe, Italy began instituting quarantine orders on March 8th in affected regions in the North and soon after throughout the country. The first case of COVID-19 in the United States was confirmed on January 20th, 2020 while the first known case of community-spread was confirmed on February 23rd. Washington was the first state to issue a state of emergency on February 29, with New York and California following the next week. Many states banned large gatherings and closed schools over the next few weeks. California was the first to issue a state-wide, stay-at-home order on March 19, asking all those not engaged in essential activities to stay at home. In the next 3 days, New York, New Jersey, and Illinois issued similar directives. By March 30, 29 more states had similar directives. But not all states implemented such restrictions: Arkansas, Iowa, Nebraska, North Dakota and South Dakota had held off on imposing stay-at-home orders, favoring other restrictions, as of April 20, 2020 (see here).
- Online job vacancy postings began to collapse in the middle of March, at about the same time as initial unemployment claims spiked. We observe a 30 percent decline in online job vacancies compared to the level at the beginning of the year using data obtained from Burning Glass Technologies (BGT), a company that scrapes, cleans, and codes job vacancies posted on the internet at a daily frequency. To put this number in perspective, the observed decline in vacancies posted, which took place between mid-March and the second week of April, is roughly two-thirds of that observed in the U.S. during the Great Recession, when a comparable measure of vacancies declined by about 50 percent over a period of 1.5 years.
- The labor market collapsed at the same time across the U.S. irrespective of the early spread of the virus or the timing of state-level policies imposed. There is little evidence that labor markets in states that were hit more heavily by the epidemic or that imposed stay-at-home orders earlier were differentially affected. Across all states, job postings decline by about the same magnitude from the middle of March onwards, regardless of whether states were earlier to implement stay-at-home orders (before March 31), later, or had not implemented such orders at all (see chart). Unemployment insurance claims also begin to rise at the same time across states, irrespective of when (or whether) the states issued stay-at-home orders. For UI claims there are small differences in the magnitude of the increase in postings across state groups that does broadly correspond with the timing of state policies: states that implemented stay-at-home orders first see a somewhat larger increase in unemployment claims. However, these differences pale in comparison to the overall claims increases. The fact that employers began to lay off workers or reduced their job openings before they had a local surge in cases or stay-at-home orders likely indicates that they were responding to a changing outlook and increased uncertainty that impacted the whole country. Based on both UI claims and job postings, we conclude that the collapse of the labor market was sudden and severe and that it displayed a high degree of synchronicity across states.
- The labor market is in broad retreat across almost all industries, whether they are deemed essential or non-essential. Using the job posting data we are able to separate out “front line” jobs such as those in essential retail and nursing. We also show patterns separately for health industries, and other types of essential industries (using New York Governor Andrew Cuomo's guidance as a reference for the classification). In the three weeks following March 14, postings increased substantially in the essential retail sector, which includes grocery and pharmacy workers among others, and remained at their pre-period level in nursing. But these types of positions were the exception. All other industries display rapid declines in postings. The steepest such decline is observed for non-essential retail, which sat below 50 percent of its beginning-of-year postings level as of April 11. However, we also observe declines in other essential and non-essential industries, each of which sit between 60 and 70 percent of their beginning-of-year postings level. Indeed, the decline in other essential industries (such as restaurants, utilities, and transportation) exceeds that in non-essential industries outside of retail.
- Unemployment insurance claims in Washington State spiked across all industries. In Washington, one of the few states that provide unemployment claim information broken down by industry, we find claims from non-essential industries rose the most: by roughly 10% of industry employment. Claims from essential retail and nursing rose the least, but still saw substantial spikes of over 2 percent of their respective employment levels. Initial unemployment claims from healthcare and other essential industries each rose by about 5 percent of their respective employment levels. So despite meaningful differences across industries, overall, UI claims show broad-based increases in Washington, including essential and non-essential workers, as well as those in healthcare.
- Job vacancies collapsed across occupations, regardless of whether work from home is possible. Using a measure of work from home capability (defined based on the tasks performed in the occupation), we find that occupations that can likely be performed from home actually saw a slightly larger drop in postings than occupations that cannot be performed from home. This is true within both essential and non-essential industries. It is not possible to differentiate the rise in unemployment claims for positions that can be performed from home from those that cannot at the national level. But, unemployment insurance claims from occupations where work from home is possible saw a smaller spike in Washington State. Claims rose by roughly 3 percent of employment in work-from-home occupations and by 7 percent of employment in non-work from home occupations. Occupations that are work-from-home capable may be insulated from some of the labor supply constraints imposed by stay-at-home orders.
What this Means:
The labor market is currently experiencing unprecedented weakness. We observe broad-based declines in job postings and spikes in unemployment insurance claims across states, sectors, and occupations. Essential retail and nursing are exceptions in that job postings have not fallen, even while they have seen large spikes in claims. As a result, these areas will see a great deal of churn, with many workers being sent to unemployment even while employers desire to maintain hiring. Work-from-home occupations have been somewhat sheltered from immediate unemployment, exhibiting less than half the rate of increase as non-work-from-home occupations. However, given the large decline in job postings for work-from-home occupations, we expect these areas to show weakness in excess of that indicated by unemployment insurance claims in the coming months. That is, these areas still suffer from the negative aggregate demand shock stemming from the crisis. The broad based nature of the collapse in vacancy postings and spikes in UI claims suggest the current damage is not solely caused by stay-at-home orders; it is too large and pervasive. Just as employers across the nation reduced hiring and began to lay off workers at about the same time — regardless of the specific stay-at-home orders in their state — lifting the stay-at-home orders does not necessarily or automatically change the outlook for demand of products and services. Customers are likely to remain cautious, companies may be reluctant to act too quickly, and interruptions in global or national supply chains are likely to continue while the risks of viral outbreaks remain. We therefore conclude that the damage to the economy is unlikely to be undone simply by lifting stay-at-home orders.