Discover more from Joshua Gans' Newsletter
Cities and Covid Reopening
Different cities have distinct network structures and so required tailored policies
Welcome to Plugging the Gap (my email newsletter mainly about Covid-19 and its economics). My goal is for several posts a week explaining economic research and the economic approach to understanding the pandemic. I may also deviate to other topics as well. (In case you don’t know me, I’m an economist and professor at the University of Toronto. I have written lots of books including most recently on Covid-19. You can follow me on twitter (@joshgans) or subscribe to this email newsletter at the link below).
Today’s post is about reopening and the value of locality specific data in tailoring government policies and recommendations.
As governments struggle to manage the ebb and flow of Covid-19, we are seeing varying responses amidst an increasing corpus of scientific information directed at helping those responses. It looks like most places will cycle between closing and opening for the foreseeable future. And inconsistencies arise. In many places, bars reopened prior to schools which it is puzzling since schools appear to dominant bars in priority on the two criteria of economic value and health safety. The same is true within education: why are colleges opening rather than schools when it should be the other way around?
In Melbourne, which this week launched back into lockdown, mask-wearing is mandatory both inside and out. This is puzzling as there is little evidence that that the coronavirus spreads outside suggesting that what you want to do is get people outside. Having to unnecessarily wear a mask outside couple with Melbourne being at the height of winter surely is the exact opposite of what you want to do. If I had to guess someone told the Victorian government that “if you don’t tell Australians to wear masks all of the time, they won’t understand what to do.” Here in Canada, we have dealt with the, surely easy to understand, concept of different policies inside versus outside just fine with respect to masks. Then again, as an Australian, I can also attest that it took some time for me to work out the notion of changing clothes inside versus out in dealing with the winter. Nonetheless, some small nuance in the policy seems warranted.
With respect to nuanced policies and, in particular, tailoring, some recent economic research suggests there are big gains to this. The goal in lockdowns is not to lockdown per se but to restrict activities in a way that you will contain the viral spread in the general population while at the same time imposing the minimal economic cost. If every location was similar, we could have a set of national guidelines that was followed and that would be that. But what recent research has found is that different localities are very different indeed.
Let’s jump right in that the deep end. This paper (by a network of authors, Mohammad Akbarpour, Cody Cook, Aude Marzuoli, Simon Mongey, Abhishek Nagaraj, Matteo Saccarola, Pietro Tebaldi, Shoshana Vasserman, and Hanbin Yang) looks at the network structure of cities and how different policies might alter that structure. Here is their starting point:
The above diagrams show hypothetical links between children (green), manufacturing workers (red) and service workers (black). If you keep firms and schools open, there are lots of links between them and everyone is connected. Shut schools and the manufacturing plant down, and you sever lots of those links but not all. A full lockdown divides the economy into the smallest number of components (that is, groups of people who are connected with each other but not others). And then there is work from home (for the service firms), shut down manufacturing and shut down schools. You end up with more connections than a complete lockdown but still much fewer components and those who are connected are connected by a single link. There is an argument that option (d) is the best compromise between economic value and healthy safety and that there is little to be gained on the latter by going to (c).
That’s the theory. But what the paper shows is that different US cities really differ in their networks. They build a model that takes into account age (including the different health impacts of Covid-19 on age), work patterns, commuting patterns, density and the ability to work from home. Here is their main finding:
This figure compares the employment versus expected death outcomes for various policies anchored to a cautious reopen (the policies most places are pursuing in reopening which limits interactions but allows people to go about their business). The policies are cautious reopening (CR), essential only (EO), work from home if possible (WFH), isolate those over 60 (60+) and alternating schedules (AS) where students and workers are split into two groups who don’t interact and alternate days coming in. Notice that the trade-offs between these policies differ between cities. A policy of work from home generally outperforms both 60+ and CR on both key dimensions. In Chicago and New York, moving from WFH to AS to EO reduces deaths but at the expense of higher unemployment. But in Sacramento, there is no health safety benefit and WFH is unambiguously the best policy.
What should we make of this? First, there are big returns for gathering the data necessary to understand city network structure. For instance, if Sacramento had used New York data and believed that it would reduce deaths by 10 per cent by going to EO from WFH, the model predicts they will be disappointed to learn that there was no reduction in deaths but a significant doubling of employment losses.
Second, understanding these issues does not require ongoing surveillance. The network structure and demographics of cities are pretty stable over time. Thus, you need only investigate once in a while to get enough information that you could assess trade-offs of different policies over time as the prevalence of the disease changes.
To that end, the paper authors are helping out. They are doing that work, adding cities as they go. You can find the outcomes here at The Reopen Mapping Project. And, of course, the code is open-source if you want to help.
Want to watch more on re-opening? Here is a video of a panel I participated on with Emily Oster that talked about re-opening — in particular, for schools.