COVID-19 Economic Impacts in the Twin Cities

How the pandemic affected the region's workers and households

In March 2020, businesses, households, and governments rapidly curtailed economic activities in response to the COVID-19 pandemic. Data from surveys and unemployment insurance claims paint a picture of economic collapse: nearly 50 million people filed for unemployment benefits between March and June 2020, including 495,000 in the Twin Cities region.

To put the economic dynamic in context and to promote understanding of the social and economic impacts, the Community Development research team analyzed the likely effects of the employment and income losses. Our findings showed:

  • Poverty rate changes and housing cost burden changes depended on the duration of business closures and reduced economic activity.
  • ​People of color have experienced a disproportionately large share of job losses since the beginning of the pandemic. The benefits provided by the federal relief package will temporarily reduce disparities by race and ethnicity, though disparities will remain quite large.
  • Relief from the federal relief package (the CARES Act) mostly mitigated early earnings losses. Our model did not include later federal relief packages. 
Job losses varied by industry and occupation groups

Communities of color experienced the largest economic impacts

The Federal CARES Act temporarily mitigated some of the pandemic's economic impacts

COVID-19 Impacts on Unemployment, Income, and Housing Cost Burden

COVID Impacts Presentation.png

Metropolitan Council Research designed a scenario analysis of how the Twin Cities region's workforce and households are impacted by the current economic crisis. We sought to answer: What impacts will job losses have on households' incomes? How much will federal responses mitigate the economic pain?

Our analytic approach was to summarize job loss rates from the national, monthly Current Population Survey and link these to an American Community Survey microdata sample for the Twin Cities region. We then build scenarios for the Twin Cities region: We apply national field-specific job loss rates to Twin Cities region sample; subtract estimated lost employment and lost earnings; and calculate new outcomes for workers and households, for varying lengths of economic "shutdown," with and without the federal relief response.

Results of the analysis include resulting employment, incomes and income change, poverty rates, and housing cost burden rates for each demographic group. All of these results are tabulated by occupation and industry sector, by race, by housing tenure, and by pre-crisis income quintile.

See the slides from the May 20, 2020 presentation to the Committee of the Whole (PDF)

The following is a broad summary of how we examined the impact of COVID-19-related job losses on the region.

  1. First, we used data from the U.S. Census Bureau’s Current Population Survey to develop estimated rates of job loss for different combinations of industries and occupations.

  2. Second, we simulated job losses by assigning a “COVID job loss” status to each worker in American Community Survey (ACS) microdata (a sample of about 96,000 residents in the Twin Cities region). This status is based primarily on the job-loss rate for their industry and occupation (from #1), but also includes adjustments to align our estimates with unemployment insurance filings data from the Minnesota Department of Employment and Economic Development (DEED).

  3. Third, we simulated the effect of these job losses by subtracting earned income from each newly unemployed worker’s income, and then calculating new household and family incomes. We did this under 24 scenarios.

  4. Finally, we calculated new measures of per capita income, poverty, and housing cost burden from these revised household and family incomes rates under each scenario, along with breakdowns by race/ethnicity, cultural groups, “pre-pandemic” household income ranks, tenure, industry, and occupation.

  5. To avoid the possibility that any one random assignment of job losses to our sample of workers would produce unrepresentative results, we repeated the above calculations 50 times and averaged the 50 different results.

See details about methods and sources (PDF)