Principal investigator: Eduarda Miller de Figueiredo
Article title: DETERMINANTS OF DISPARITIES IN COVID-19 JOB LOSSES
Article authors: Laura Montenovo, Xuan Jiang, Felipe Lozano Rojas, Ian M. Schmutte, Kosali I. Simon, Bruce A. Weinberg, and Coady Wing
Sample size: 43.754 individuals
Location of the intervention: U.S
SectorJob Market
Type of interventionThe effects of Covid-19 on unemployment.
Main variable of interestUnemployment
Assessment method: Others
Policy Problem
The pandemic severely reduced social and economic activity, with large sectors of the economy ceasing normal operations while the government's social distancing order was in effect. In May, in some American states, both the public and private sectors restarted the reopening of economic activity.
However, in March 2020 there was a significant increase in initial unemployment insurance claims which, regardless of state policy, demonstrated a large drop in job offers (Kahn et al., 2020). Similarly, Coibion et al. (2020) estimate that unemployment far exceeded unemployment insurance claims in early April.
Assessment Context
The literature demonstrates that existing patterns of social stratification shape socioeconomic outcomes during economic crises.
For example, the fact that the Great Recession reduced the life expectancy of older workers, particularly white men, as suggested by Dudel and Myrskylä (2017). Or also the assessments of the short- and long-term effects of Hurricane Katrina on subgroups of evacuees, from the perspective of the labor market, fertility, marriage, among others (Zissimopoulos and Karoly, 2010; Schneider and Hastings, 2015; Grossman and Slusky, 2019).
Given the particularities of the economic crisis stemming from Covid-19, it is relevant to understand which population subgroups were most affected, why, and how these effects may lead to long-term well-being disparities.
Policy Details
For this study, the authors used a sample of data from the Basic Monthly CPS from February to May 2020, which includes all workers aged 18 to 65. The database has complete information on gender, race, education, state, residence, recent unemployment status, occupation code, and the CPS weight of the sample.
The authors defined a worker as recently unemployed if the person is registered as unemployed in the month of the survey and this status has been in place for a maximum of 5 weeks in March, 10 weeks in April, and 14 weeks in May. Those defined as “absent from work” are employees who were “temporarily absent from their regular job due to illness, vacation, bad weather, labor disputes, or personal reasons, whether or not they were paid for the time off” (US Census Bureau, 2019). There was a massive increase in this type of worker during February and May 2020, deserving particular attention, as some laid-off employees may have been registered as short-term absent instead of unemployed. Thus, the authors analyze measures of recent unemployment and employed but absent separately.
Data from the Occupational Information Network (O*NET) Work Context of 2019, which reports measures for 968 occupations, were also used. This database contains several questions, such as how employees perform specific tasks, the need for face-to-face contact with clients and colleagues, and the possibility of work being performed remotely. These measures were captured before the pandemic, therefore they do not capture the innovations in work practices that were induced by the health crisis.
Methodology Details
To compare unemployment across three recessions, the authors make the following distinction:
- 2001 Recession – March to November 2001;
- Great Recession – December 2007 to June 2009;
- Covid-19, lockdown phase – February to April 2020;
- Covid-19, reopening phase – February to May 2020.
Seeking to connect work subgroups and worker characteristics, the authors ran regressions where the dependent variable indicates whether the individual is recently unemployed or temporarily absent from work. Using a vector set of covariates, the authors also employed fixed state effects, due to the specific epidemiological conditions of each state.
Results
Job losses during the first months of the Covid-19 pandemic exceeded the losses of the other two recessions, which spanned 9 months (2001 Recession) and 19 months (Great Recession), as can be seen in Graph 1. Furthermore, regressions suggest that recent unemployment rates vary with individual characteristics.
Almost no group is spared from unemployment during any of the three recessions. However, young people aged 18 to 24 and Hispanics had the worst outcomes during the Covid-19 pandemic. According to the authors, this occurred because these groups work in sectors that are particularly affected by social distancing measures. This is because Hispanics work more in jobs that cannot be done remotely, and young people have jobs with greater face-to-face interaction.
Unemployment was also polarized by education level, meaning higher unemployment among those with only a high school diploma or a bachelor's degree. More qualified workers are more likely to occupy essential positions, which justifies a lower unemployment rate.
Chart 1 – Job Changes in the Three Recessions
Regarding family structure, married individuals have a lower unemployment rate than single individuals. However, women with young children have a higher rate of "employed but absent," which is concerning and may indicate future unemployment. Single parents, mostly women (72%), suffered the most from unemployment during the pandemic. However, the data shows that the age of the children is weakly related to job changes during the crisis. Thus, efforts to support new childcare options are extremely important in this pandemic scenario.
Upon analyzing the results, the authors realized that a substantial portion of the unemployment difference between the subpopulations can be explained by differences in pre-pandemic classification between occupations that are more and less sensitive to the Covid-19 shock.
Lessons in Public Policy
In summary, the study suggests that the Covid-19 pandemic had different impacts on subpopulations, and that these impacts were distinct from those felt in recent recessions. Therefore, unemployment rates are very high among younger workers when compared to other subpopulations and also when compared to previous recessions.
Reference
MONTENOVO, Laura et al. Determinants of disparities in covid-19 job losses. National Bureau of Economic Research, 2020.
