Iowa and the coronavirus: 22 counties, 77 counties

Iowa map showing counties under separate COVID order (swiped from wgem.com)

Iowa like other states is trying to figure out how to manage the ongoing coronavirus pandemic. Governor Kim Reynolds has issued a series of proclamations, while stopping short of an all-encompassing shutdown as experienced in neighboring states like Illinois and Minnesota. Most business restrictions will expire at the end of this week, with remaining exceptions bars, casinos, and public gatherings of more than ten people (Hadish 2020b).

Prior to this, most restrictions had been lifted in most counties, but retained in 22 counties, colored orange in the map above (Roberts 2020)--which I must say, before we go any further, looks a lot like this...

...which relates either to I Thessalonians 4:16-17, or to a geometric figure with finite volume and infinite surface area. But I digress, don't I?

There is certainly room to question the timing of reopening (Abutaleb et al. 2020, Berch 2020), and even whether more counties should have been covered by the stricter rules (Iowa Fiscal Partnership 2020).The governor herself is on "modified quarantine" after interacting with President Donald Trump and Vice President Mike Pence last week (Hadish 2020a).

("I shouldn't punish half of the state when we've got a significant spike in eight areas," Reynolds said at the time (Henderson 2020). This seems to me the most bizarre formulation in a series of bizarre official communications. Protecting the public health and safety is not punishment. Unless it is?)

However, we are not going to go there. On the assumption that there is or was some qualitative distinction in virus severity between the 77 counties mostly-opened on May 1 and the 22 counties mostly-opened on May 15, we are going to soak and poke in some data to see what that distinction might be.

We begin by noting that growth in Iowa has been concentrated in a relative few of our 99 counties, mostly clustered around the cities of Cedar Rapids, Davenport, Des Moines, Dubuque and Iowa City. Updating the data from my 2018 post on the subject, the ten largest urban-suburban counties have added 142,659 residents since the 2010 census, which means the other 89 counties combined are at minus 35,000 over that same period. Linn County is up 15,463, more than the Cedar Rapids metropolitan statistical area as a whole, while the Des Moines metro area's growth has included fast-growing Dallas County to the west and Warren County to the south as well as Polk County. 

The ten urban-suburban counties now account for 52.5 percent of the state's population, but have added 84 percent of the state's new jobs in this decade, attracted 312 percent of people moving into the state, and account for well over half (56.6 percent) of people aged 25-44. [The other 89 counties have a combined net out-migration, so the number of migrants into the core 10 more than triples that of the whole state.] Polk and Johnson Counties, famously contrasted by the Governor with "the real Iowa" in 2018, while comprising 20 percent of Iowa's population, have together added 80,000 residents since 2010, along with 42.5 percent of the state's job growth, 31 percent of the state's economic output, 202 percent of people moving into the state, 26.9 percent of college graduates, and 29.3 percent of those with graduate or professional degrees. Not sure what "the real Iowa" would do without them. 

For the record, since we've been talking about density of late in relation to the pandemic, the density of the ten largest counties is 254 people per square mile, ranging from 750.5 (Polk) to 98 (Pottawattamie). Density is 54.5 for the state as a whole, 30.4 for the "other 89."

The 22 counties whose COVID situation merited a two week delay in lifting restrictions include eight of the ten largest counties, excepting Pottawattamie (Council Bluffs) and Story (Ames); three that are part of MSAs (Benton, Bremer, and Washington); and eleven that are in neither category. 

Are the 22 counties different from the 77 in some characteristic(s) that would lead us to expect the pandemic would be more severe there?

The density hypothesis (Rosenthal 2020) posits that places with greater population density will see greater community transmission of the virus, with New York City and its environs the example that springs immediately to mind. The 11 in MSAs include the state's five densest counties, as well as Dubuque, Woodbury, and Dallas at #7-8-9. Benton and Bremer, though, are at about the statewide average, and Washington County (38.2/sqmi) is quite a bit below. The 11 non-MSA counties include two in southeast Iowa with near-MSA-level density, Des Moines (96.9) and Muscatine (97.7). [NOTE FOR NON-IOWANS: Des Moines County does not contain the City of Des Moines. In fact, they are nowhere near each other.] However, five of the counties are below average for non-MSA Iowa counties, including Allamakee (22.3/sqmi); the median case, Poweshiek County, is at 32.3/sqmi almost exactly the statewide average for non-MSA counties. So, if your county is densely-populated it was slightly more likely to land on this list, but being sparsely-populated was no insurance against inclusion.

Sure is dense in here! (Reopened bar in West Allis, Wisconsin, swiped from WDJT)

The connectedness hypothesis posits that as the virus spreads, it will spread first to places that are most connected (economically, socially) to the rest of the world, and only eventually will get to more outlying areas. Iowa does not have any global cities, but there are definite variations in local economies, as shown by variations in job growth, in-migration, and percentage of graduate and professional degrees. Successful cities tend to be distinct from the rest of the state on these factors, but not all cities are successful. Weirdly, the two largest counties among the 77 display nearly opposite experiences: Story County, home to Iowa State University, is second only to Johnson in graduate and professional degree holders at 2.5 times the statewide average; it is 6th in in-migration and 7th in GDP. And it's right by Des Moines (the city). Pottawattamie is comparable to Story in GDP, but for advanced degrees has only 2/3 the statewide average, and has had substantial out-migration (though less than Black Hawk and Woodbury Counties, which are among the 22). It's also right by the big city of Omaha, Nebraska. Among the 11 non-MSA counties, only Jasper (+38) had net in-migration in the 2010s; the relatively-dense Des Moines and Muscatine Counties also had substantially larger GDP; and all are at or below the statewide average for advanced degree holders except for Powesheik, which is home to Grinnell College. These data are no more conclusive than the density measure! and there's a lot of coincidence between the two.

Besides the lack of conclusiveness on these two dimensions, there are just too many confounding factors to draw a solid conclusion on what areas are most at risk for spreading the coronavirus. A lot has to do with local incidence of things like meatpacking plants and nursing homes, how people behave in specific areas, and maybe even good or bad luck. The situation in late April when the Governor made her decision was a single snapshot in an evolving event; whatever data she and her advisers were using, they might have produced a different set of counties two weeks earlier or later. And there's the further confounding factor that the data on coronavirus incidence, hospitalizations, and deaths are shaky and likely to be understated (see Kristof 2020).

In the two weeks after the Governor's proclamation in question, 22 Iowa counties attained High COVID-19 Prevalence as defined by William H. Frey of the Brookings Institution i.e. 100 cases per 100,000 population. From April 27 to May 3: Crawford, Dubuque, Fayette, Greene, Grundy, Guthrie, Howard, Jones, Lyon, Monona, Shelby, and Wapello. From May 4 to May 10: Audubon, Boone, Buchanan, Buena Vista, Clayton, Davis, Des Moines, O'Brien, Plymouth, and Sioux. Only Dubuque and Des Moines counties were on the Governor's list of 22. (Five other counties not on the list--Cedar, Clinton, Harrison, Osceola, and Van Buren--had attained High COVID-19 Prevalence before April 27.) Frey sees the virus moving into "Trump counties;" 20 of the 22 new counties were carried by Reynolds in 2018.

Perhaps the most reasonable conclusion from all this is that the coronavirus does not have opinions about partisan politics, or walkable urbanism, or the state of the economy, and it should be treated as a live problem wherever you are. Ask not where the coronavirus is coming... it's coming at you.

EARLIER POSTS

"Election 2018 and What Happens Next," 20 November 2018
"What is the Future of Iowa's Small Towns?" 3 July 2018