While the worst effects of the COVID-19 outbreak in the United States have been relatively concentrated in a handful of cities and states, early indicators suggest that the economic fallout of the virus will be widespread.
Weekly data on unemployment insurance claims from the Department of Labor have shown a staggering increase since the outbreak began, with over 22 million claims in the last four weeks. In this piece, I’ll dive into the data a little further, investigating which states are being hit the hardest economically, and whether that lines up with the states being hit hardest by the virus itself.
Dashboard 1 below begins to answer these questions. The top graph in the dashboard shows the number of confirmed cases as of 4/16, adjusted for each state’s population. It shows that the number of cases drops off rapidly, from New York (11.4 confirmed cases per 1,000 people), to New Jersey (8.5), to Louisiana, Massachusetts, and Connecticut (4 – 5 each).
The bottom graph, which shows the number of unemployment insurance claims filed in the weeks following the outbreak of the virus (3/15 – 4/4), tells a different story. This graph shows that the economic effects of the virus are being felt in a wide variety of states. While some of the hardest hit states economically are also among those with higher rates of confirmed cases (e.g. Rhode Island, Michigan, and Pennsylvania), it’s clear that plenty of other states are experiencing widespread economic pain as well.
The same data is also available in the interactive map in Dashboard 2. You can select which of the two metrics you’d like to see in the dropdown menu on the right.
To investigate the potential relationship between confirmed cases and unemployment claims further, Dashboard 3 plots the two metrics against each other. The trendline indicates that there could be a slight positive relationship between the two, though it’s not statistically significant at any conventional significance level (p-value of 0.18). However, it’s worth noting that if you exclude New York, the relationship becomes significantly more positive and is statistically significant at a 95% level (p-value of 0.03). The dashboard is interactive, so you’re welcome to exclude any given state(s) by clicking on the data point and selecting the “Exclude” option. It’s also worth noting here that with new data on COVID-19 infections being released daily, and data on unemployment insurance claims being released weekly, this picture is likely to continue evolving over the coming weeks.
Regardless of whether such a relationship exists, what these dashboards make clear is that there will be widespread economic fallout as a result of the COVID-19 pandemic. While the CARES Act was a good first step – including a meaningful expansion of unemployment insurance benefits in addition to the $1,200 stimulus checks and other measures – it’s clear that Congress will need to take further action to cushion the economic pain Americans are experiencing.
Finally, it’s worth addressing how this analysis should not be interpreted. As the economic and psychological effects of social distancing and stay-at-home orders continue to grow, we’re likely to see an increasing chorus of calls to “reopen the economy” and stop social distancing. This is already starting to play out as thousands of people are taking part in protests against stay-at-home orders across the country. This idea of “reopening the economy” will likely be especially tempting in states and areas that are not currently experiencing widespread infection, but are suffering economic consequences nonetheless.
However, it must be stated that this would be a gravely dangerous interpretation of the data. There is good reason to believe that social distancing is working, and experts warn that it’s still too early to relax stay-at-home orders. It’s clear that until we’re able to massively ramp up testing and/or put in place the infrastructure to manage surveillance and contact tracing on a mass scale, efforts to relax social distancing restrictions are extremely risky, and the “choice” between saving the economy and saving lives is a false one. The fact that we do not appear to be close to meeting the levels of testing or contact tracing needed to relax social distancing regulations only further underscores the need to take aggressive action to address the growing economic pain Americans are facing.
You can find the data and code used in this analysis here.
