COVID-19 Economic Crisis: By State

View these graphs on Carsey's Tableau page

Key Findings

icon united states

Every state has lost more jobs than it lost in the Great Recession.

Icon of a group of people

More than 15% of jobs were lost in 17 states from February to April.

icon coronavirus

Of the 10 states that lost the most jobs, 7 were in the top 10 in number of COVID-19 cases per-capita. 

icon up arrow

In every state lower wage industries lost far more jobs than high wage industries.

Job loss is sweeping the country. As of mid-April, the latest period for which state-by-state official data are available, every state in the country had seen job loss compared to February. Nationally, there was a rebound in May but, nevertheless, by mid-May close to 20 million jobs had been lost since February—a loss of 13% of all payroll jobs in the country. As of mid-April job loss in states ranged from Michigan, with 23.5% of payroll jobs lost (over a million jobs) to Oklahoma with 8.5% (145,000 lost jobs). Almost every industry has felt the impact in nearly every state. Over half (56.7%) of jobs in the Accommodation and Food Services industry were lost in Michigan. But the state lost 11.8% of its Wholesale Trade jobs as well. The maps, graphs, and tables that follow show the story.

Map 1 compares states in the percent change in the number of jobs between the Bureau of Labor Statistics (BLS) February survey and the April survey (covering payroll periods containing April 12). Seventeen states lost over 15% of their jobs. The states losing the highest percentage of jobs were Michigan, Vermont, and New York. For comparison, 12 states showed only small overall job loss between January to February of this year.

The strongest determinant of the number of jobs lost as of the April BLS survey was how hard COVID-19 had hit the state. Of the 10 states that lost the most jobs, 7 were in the top ten in number of cases per-capita at the time of the BLS survey. The other three were Vermont, Hawaii, and Nevada which rely disproportionately on the most impacted industries—particularly in Leisure and Hospitality. Michigan topped the list largely because, in addition to ranking 8th in cases per-capita, its manufacturing sector was hard hit even relative to other states with high numbers of COVID-19 cases. 

  1. Employment data used here are seasonally adjusted, meaning that the Bureau of Labor Statistics accounts for normal seasonal variation in employment due to various factors.
  2. Small differences between states in the percent job change may not be statistically significant.

In the hardest-hit Leisure and Hospitality Sector, New York had the largest percent drop at 66.5%—the overall national decline was 48.3% by April. The sector’s component industries both saw sharp declines in every state. Accommodation and Food Services declined the most in New York, 67.7%—with the national drop being 47.3% by April. Arts, Entertainment, and Recreation jobs fell the most in New Jersey where 73.4%, or 54,800, of those jobs were lost—nationwide, there was a 54.5% decline into April.

After Leisure and Hospitality, the Other Services sector was most impacted with a 22% national job loss by April. The state with the biggest percent loss in this sector was Massachusetts at 37.9%. Nationally, the industry that contributed most to this sector’s decline was Personal and Laundry Services—although there were also large losses in Repair and Maintenance and in Religious, Grantmaking, Civil, Professional, and Similar Organizations (industry-level data are not available for states in this sector).

The next hardest hit industry was Administrative and Support and Waste Management and Remediation Services, which is in the Professional and Business Services sector. Nationally, the decline was 17.3% into April, with Michigan seeing the largest drop at 34.4%. Most of the jobs in this area are in temporary help, placement agencies, and services provided to buildings.

Michigan was also hardest hit in Manufacturing with 28.7% of jobs lost—a total of 179,100 jobs. Government employment also fell across the country. Wisconsin saw state government employment drop by 31.6%. Connecticut saw a reduction in local government employment of 14.6%.

Map 1 can display individual sectors and industries as well as employment for all sectors and industries. Selecting a state in the map shows the percent and number change in jobs for the state. The table in the addendum shows a top-to-bottom ranking, sortable by percentage change, total employment change, and state name. The table can list employment in all sectors and industries or individual sectors and industries. Descriptions of each sector and industry can be found on the BLS website.

Chart 1 compares (1) the change in number of jobs during the Great Recession (December 2007 to June 2009), (2) job loss in the current crisis through mid-April (February to April), (3) job change in March, and (4) job change in April. Comparisons can be viewed for the nation or a selected state, district, or territory as well as for total payroll employment or a selected sector or industry. In total employment, every state has already lost more jobs than it lost from its peak employment to its low point in the Great Recession. Every state has lost more jobs than during the Great Recession in Accommodation and Food Service, 46 have lost more in Retail, 42 have lost more government jobs, and 46 more in Health Care and Social Assistance.

  1. Employment data are seasonally adjusted.

Chart 2 shows in stark fashion how catastrophic the job loss has been. Between February and April, in every state, essentially every job gained since the Great Recession was lost. This is a sharp reversal from the steady climb that began in 2010, accelerated starting in 2012, and had slackened in the last few years.

The causes of the Great Recession and the current crisis are very different, and the paths to recovery will be different. In this crisis, the industries that have been hardest hit are unlikely to fully recover quickly. Although there will be an initial rebound as some formal restrictions are lifted, continuing restrictions and an understandable ongoing reluctance to enter crowded establishments as long as the public health risk remains will continue to slow recovery. In addition, employers will be weakened or have ceased operations, and it will take some time to fully recover. The upward line to come will not be as pronounced as the downward line so visible in Chart 2. How pronounced it will be is the critical economic question.

  1. Employment data are seasonally adjusted.

Chart 3 shows the changes in employment for the selected jurisdiction for each industry (or for some states, sector). In every state Accommodation and Food Services has contributed the most to job loss. Health Care and Social Assistance has also been hard hit in every state. Within this industry losses nationally have been mostly in Ambulatory Health Care Services, including physician and dentist offices. Hospitals, residential care, and Social Assistance also saw significant declines. Included in Social Assistance is Child Day Care, which saw a drop in employment nationally of 34% by April.

  1. “SA” indicates that the data for the sector or industry are seasonally adjusted. “NSA” indicates that the data are not seasonally adjusted.

Chart 4 shows the relationship between job loss and average weekly wages for each industry. The overwhelming number of jobs has been lost in low-paying industries. Nationally, Accommodations and Food Service had lost over 6.8 million jobs (a 47% decline) by April and had an average weekly pay of $436 in the third quarter of 2019 (the latest period for which data are generally available). Retail, and Health Care and Social Assistance, each had lost about 2.2 million jobs and had weekly pay nationally of $640 and $988, respectively. Meanwhile, industries such as Information, Management of Companies and Enterprises, and Finance and Insurance each had lost less than 500,000 jobs and had average weekly pay of $2,221, $2,127, and $1,812, respectively. These are, of course, averages across a wide range of jobs within the industries. Finance and Insurance includes both the heads of Wall Street firms and bank tellers. It is clear, however, that job loss has been significantly more acute in low-wage jobs than high-wage jobs. Recent Census Bureau data show that there have been more workers with employment income decline in low-income than high-income households.

Low-wage employees are obviously much more heavily impacted by job loss. They have less in savings, and those who had employer-provided health insurance are less likely to be able to afford continuing coverage.

  1. Hourly pay levels are not available as comprehensively as weekly wages at the state level so they are not used here, but the pattern for hourly wages is similar to the pattern for weekly wages.

Table 1 shows the percentages of job loss for each industry in the country, selected state, district, or territory. Industries with more employees are, of course, more likely to contribute more greatly to the total numbers of job losses shown in Charts 3 and 4—just as larger states contribute the most to national job loss. From the perspective of an employee, customer, owner, or other stakeholder in an industry, the level of impact is measured best by the percent of jobs in the industry that have been lost. It is notable that, for the most part, the reason the most jobs have been lost in low-wage industries and the fewest in high-wage industries (as highlighted by Chart 4) is not just that low-wage industries employed more people prior to the current crisis. Higher paying industries, while certainly feeling an impact, have seen a much smaller overall share of job loss. A waiter, for example, is much more likely to have lost his or her job than a banker.

  1. “SA” indicates that the data for the sector or industry are seasonally adjusted. “NSA” indicates that the data are not seasonally adjusted.

Unemployment Rate

Map 2 shows the unemployment rate. This measure is likely to be somewhat unreliable for the coming months. It is a measure of those who are not employed as a share of those who are employed or seeking employment. How respondents respond to survey questions related to whether they are seeking employment in this unusual period will not be consistent, which will make the unemployment rate somewhat unpredictable. There are other measures of unemployment, but they are not available by state on a monthly basis.

The unemployment rate was 9.8 percentage points higher nationally in May than it had been in February after being 11.2 percentage points higher as of April. The rate hit 28.2% in Nevada in April and 22.7% in Michigan. Clicking on a state in the map will show the February and April rates. The chart in Addendum 3 plots the unemployment rate from 2008 through April for the country, selected state, district, or territory.

Employment-to-Population Ratio

Chart 5 shows the Employment-to-Population Ratio (EPOP) so far in 2020. This is the percent of the population that is gainfully employed. For the country, and most states, it dropped significantly in March and April. In the long run this measure is influenced by factors such as the aging of the population out of prime working years. The dramatic economic impact of COVID-19 will, however, dominate this measure in the coming months. Nationally, almost 10% fewer of the population were employed in April than February.

Unemployment Insurance Initial Claims


The number of initial Unemployment Insurance (UI) Claims filed is the most up-to-date measure of the extent of the employment loss across the country. It is, however, a gross measure—it reflects only those losing their jobs without being offset by those becoming employed. The distinction can be seen clearly in the most recent national employment data. Employment went up in May nationally although UI claims were still at historically high levels during that period. 

Chart UI-1 shows the trend in UI initial claims for the country or selected state, district, or territory. The declining number of claims seen across the country is only mildly encouraging. These are still extraordinarily high levels of unemployment claims—and 3 states saw their number of claims rise last week. For comparison, the chart also shows the highest weeks of UI claims in 2008 and 2009 during the Great Recession. Initial unemployment claims have been higher in last 11 weeks in 27 of 50 states than in the worst weeks of the Great Recession in 2008 and 2009. Variation from week-to-week can reflect the pace at which claims are handled—states with systems that have been overwhelmed are less likely to see a drop off even if there are fewer newly unemployed.

Map UI-1 shows where UI initial claims are being filed as a percentage of the February 2020 labor force. The map allows for selecting for total claims since March 7, April 18, or latest week (the week ending May 30). The table in Addendum 2 has the sortable top-to-bottom ranking and additional data by week. Differences between states in the number of claims filed is not a reliable indicator of differences in the numbers of jobs lost. In addition to the number of employees who have lost their jobs, differences can be attributable to differences in eligibility rules, filing ease, and states' efficiency in processing claims—particularly during the current unprecedented surge.

View Addendum 1: Sortable Jobs Table

The table in Addendum 1 can be sorted by percent change in employment, change in the number of jobs, or alphabetically by state. Different sectors and industry can be selected.

View Addendum 2: Sortable UI Claims Table

The table in Addendum 2 shows, for each state, total initial UI claims since the week ending March 14 and the claims as a percent of the February labor force. It is sortable by the percent, total numbers, or state name.

View Addendum 3: Unemployment Rate 2007 – April 2020

Addendum 3 shows the unemployment rate by month from January 2007 to April 2020, the most current month for which data are available. The national total or individual state, district, or territory can be selected.