What a difference a week makes! And not for the better.
In one week from March 29 - April 4, the U.S. coronavirus mortality rate quadrupled from a cumulative total of 7 to 28 deaths per million U.S. residents. The number of new unemployment claimants has gone from over 10 million to nearly 17 million jobless. Claimants in the last three weeks now equate to about 10.4% of the normalized base of 145+ million U.S. workers covered by unemployment insurance.
This blog post updates statistical information posted on April 4 regarding the state-by-state relationship between COVID-19 and joblessness. And this week, our update also looks at the breathtaking changes in both metrics as have occurred in just one week’s time.
Note: For this review, information is as of an April 4 weekly report (the latest for unemployment claims). In contrast, COVID-19 death data is posted daily. As of April 11, the nationwide COVID-19 mortality rate had jumped from 28 deaths per million population (on April 4) to over 65 deaths (April 11), led by dramatic increases in New York mortality (as shown by the map on our website home page). This linkage review will be updated for April 11 jobless data when this updated employment data is next made available Thursday, April 16.
EARLY APRIL UPDATE
The following graph compares cumulative unemployment claims filed during the weeks ending March 21, March 28 and April 4 with COVID-19 mortality as of April 4.
As with last week’s chart, cumulative unemployment claims over a 3-week period are calculated as a % of total covered employment, by state. Mortality is calculated in terms of the number of COVID-19 related deaths per million residents, by state.
If you were to compare this week’s scatterplot with that of last week, the most obvious change is that the variability of outcomes has greatly widened. As of March 29, New York’s cumulative mortality rate was 50 coronavirus deaths per million residents. One week later, the mortality rate had more than quadrupled — to about 214 deaths per million population as of April 4 (a number that has again more than doubled to 482 deaths per million as of April 11).
And the span of joblessness has widened. As of March 28, Pennsylvania and Rhode Island had experienced unemployment claims equaling 13.8% of total base employment. As of April 4, Rhode Island’s rate has jumped to 19.4%.
Several other items are of note in with this update:
There is much greater spread of mortality rates than before. In addition to New York, the states of New Jersey and Louisiana have mortality rates at just over 100 deaths per million residents. Michigan now tops 50. All other states and territories have mortality rates of less than 50 deaths per million.
There is also a much wider range of unemployment claims — from 4% to nearly 20% of the total workforce covered by unemployment insurance. At the upper end of the spectrum are Rhode Island, Michigan, Pennsylvania, Nevada and Hawaii, all in the 16-20% range. The lowest rate is noted for South Dakota at 4%. Despite having almost off the chart mortality, New York’s 8.4% rate of unemployment claims is below the nationwide 3-week cumulative rate of 10.4%.
And there is now a circled group of about 25 states (with Washington D.C. included) that so far have experienced a combination of relatively low mortality (below 45 deaths per million population) and low rates of unemployment claims (below 10%). This group includes some very rural as well as urban states - with strong representation from the sunbelt and mid-America. These are the states that, so far, might be considered as either best practice examples or just plain lucky as compared to the other half of states.
Focusing more on what is working well for half the states may help in determining a path out for further curve flattening and for economic recovery. But before jumping too quickly to conclusions, we shift the analysis to also consider changes of the last week - both in mortality and joblessness.
CHANGES IN COVID-19 DEATHS VS JOB LAYOFFS
An added feature of this update involves comparing outcomes of the period ending March 28 with April 4, one week later. The following graph displays much of the same information as the prior illustrative depicting trends in coronavirus morality versus job layoffs. The difference is that the prior graph showed the pattern of deaths versus layoffs as of a single point in time (April 4) — while the following graph shows the change from one week to the next (from March 28/29 - April 4).
As shown by the graph:
Just 8 states experienced an increase of COVID-19 mortality of at least 20 deaths per million residents in the most recent week ending April 4. Led by New York (with a week-over-week increase of nearly 165 deaths per million, other places with mortality increases of 20 or more per million residents were New Jersey, Louisiana, Michigan, Connecticut, the District of Columbia, Massachusetts and Rhode Island. The increase in mortality rates nationwide averaged 21 deaths per million residents.
All the remaining states experienced mortality rate increases of 20 added deaths per million residents or less. The extent to which these slower rates of increased mortality is due to good public policy, lower population density, and/or just plain luck varies widely across these states.
There also is wide disparity in the job layoff patterns across these states with slower growth in mortality rates. Georgia’s unemployment claims increased by 9% of the employed work force in the one week ended April 4. At the other end of the spectrum, Colorado’s unemployment claims increased by less than 2% points. In effect, there appears to be no guaranteed employment reward for those states that have contained the death toll, to date.
A PATH TO ECONOMIC RECOVERY?
At long last, serious discussion about re-starting the American economy is beginning to get underway. At this point, there is as yet no certainty as to the appropriate mechanisms or timeline for Americans going back to work. This is likely to be a subject of intense debate from health care and economic perspectives, not to mention anticipated differences across the political spectrum
Are there any observations from this coronavirus and joblessness discussion worth considering as part of this emerging policy debate? Three observations are suggested:
First, as has been previously suggested, this updated review further reinforces the observation that national and state-level policy should not be framed around the assumption that “one size fits all.” New York’s experience (together with that of adjoining states) is well beyond the pale of what has been or is likely to be experienced across much of the rest of the country. This is made abundantly clear by the very different curve flattening trajectory that the nation’s most populous state — California — has followed. And the experience of a California is yet different from that of a Wyoming or South Dakota — where mortality rates increased at 0 and 1 deaths per million residents over this most recent week.
Second, this analysis does suggest that continuous monitoring of changing mortality may serve as a useful guidepost for determining which states are best positioned to ease away form their business shutdowns — and how quickly. States like New York, New Jersey, Louisiana, and Michigan appear to warrant continued lockdown until the mortality growth rates ease back to the national average (or to below zero). Other places like DC, Connecticut and Pennsylvania that have been relatively calm so far may be erupting as data over the next one to two weeks may demonstrate — warranting more intensive measures at least temporarily.
Conversely, states that have consistently held mortality increases to well below the national average gain of 21 deaths per million this last week (to a gain of, say, no more than 10 deaths per added million), appear to be the best candidates right now for potential relaxation of shutdown requirements. Of 52 states and territories, 36 appear to be in this category right now. Of these, 20 are below a mortality gain of 5 per million for the week ended April 4.
As the country reaches the peak of mortality in the days or weeks ahead, the weekly change in national COViD-19 should go to zero and then negative. As the weekly mortality threshold for America declines, states remaining below the national curve offer the best case to be the most quickly rewarded with economic re-boot. This type of monitoring and economic adjustment may be required not only in the weeks ahead but over an extended period to limit the risk of re-infections longer term.
Third, a similar but more nuanced approach might be taken to states experiencing unemployment claims well above the national average. To what extent are higher rates due to earlier shutdowns, different industry mix (as with a high proportion of businesses dependent on face-to-face contact), unusually draconian business closure requirements, or other factors? Encourage states with unduly high rates but without demonstrable mortality benefit to begin opening sectors posing the least risks of virus transmission.
