Here’s a letter to a Café Hayek commenter:
After reading my post on nationwide hospital utilization rates in the U.S.  – a post showing that the rate today, in 2020, is not out of the ordinary – you comment:
Now do the state level of the same data. It won’t look as pretty Don. Simpson’s paradox got you. Be better.
You correctly understand that slicing and dicing data in one way can give a very different picture than is revealed when the same data are sliced and diced differently. So let’s indeed look at data on hospital utilization at the state level (including Washington, DC).
State-level data on inpatient occupancy of hospital beds are available for downloading from the U.S. Department of Health & Human Services . Here’s what these data show for the 31 days of November 4th through December 4th. Note that these are estimates. Each daily estimate for each state is offered with an upper bound, a lower bound, and a ‘best’ estimate of the utilization rate.
From November 4th through December 4th the highest upper-bound estimated bed-utilization rate is 98.98 for Maryland on December 4th. The lower-bound estimate for that same day in Maryland is 69.6%, with the ‘best’ estimate for that day being 84.29%. Assuming, reasonably, that the ‘best’ estimate is more accurate than either the upper- or lower-bound estimate, on December 4th more than 15 percent of hospital beds in Maryland remained unoccupied.
In the month spanning November 4th through December 4th, on only three days did any state have a ‘best’ estimate of hospital utilization rate of 90 percent or higher. In all three cases that state is Rhode Island, with 90.27% of its hospital beds occupied on November 12th; 90.37% occupied on November 24th; and 91.24% of its hospital beds occupied on December 3rd.
Further, during this same 31-day period, out of 1581 daily observations from all 50 states and DC, only 102 of these ‘best’ estimates of hospital utilization show utilization of 80 percent or higher (which include the three above-mentioned days in Rhode Island).*
The bottom line is that, when broken down to the state level – at least for the period November 4th through December 4th – there is no evidence that hospitals in the U.S. are close to running out of beds for patients.
Donald J. Boudreaux
Professor of Economics
Martha and Nelson Getchell Chair for the Study of Free Market Capitalism at the Mercatus Center
George Mason University
Fairfax, VA 22030
* I counted these data by hand. I did so twice, and believe my count to be accurate. But it’s possible that I’m off by a small amount – but not enough to dilute my letter’s conclusion.