In 2017, I moved from Middlesex County, Massachusetts to Orange County, CA. The two counties are somewhat similar. Both are middle to upper middle class suburban counties, full of professionals, located just outside major cities. Orange County has 3.2 million people, while Middlesex County has 1.6 million. No big deal, right?
Actually, it’s a huge deal. So far, 221 have died in Middlesex County vs. 22 in Orange County. That means the risk is 20 times higher in Middlesex County. Read that again. I didn’t say 20% higher; I said 20X. That’s a lot! I dodged a bullet.
Does that vast different in risk make any difference? Probably not. When I go out to get take out food or to go grocery shopping, I wear a mask and try not to walk close to other people. That’s almost exactly what I’d be doing if I still lived in Middlesex County. And I imagine most other people are the same.
I recall reading of a behavioral economics paper that showed that most people respond roughly equally to risk factors of 10%, 1%, 0.1% etc. [If someone knows of the paper, send me a link and I’ll add it here.]
So even though my rational brain knows that Orange County is pretty safe and I know that 20X is a really big deal; I act on my gut, which treats the two places as equally risky.
Here’s a second example:
About 34,000 Americans have already died of Covid-19 and the experts tell us that we are roughly half way through the first wave of the epidemic, which is expected to fall back this summer (and perhaps rise again next winter.) So if the models are correct, Orange County might end up with another 22 more coronavirus fatalities by late summer.
But that data point is extremely misleading, as deaths typically follow with at least a 2 or 3-week lag on first getting infected. Given the bell-shaped curve of infections, that means a pretty large share of the extra 22 deaths we can expect in Orange County have already been infected. Thus the risk going forward of going outside and getting newly infected and dying is actually quite low in Orange County, at least this spring and summer. (Of course don’t take this as license to go out and act recklessly, remember the Lucas Critique!)
In an earlier post, I argued that in the early stages of the epidemic we underreacted, and later on we would eventually overreact, at least from an individual self-interest perspective. While the risk today in America is probably lower than many people suspect, back in February and early March it was far higher than people realized. Recall Mardi Gras, the Biogen conference, and those Chinatown parades.
This overreaction from an individual perspective is actually good news from a social perspective, as there are externalities from socializing during an epidemic. That’s why it’s not obvious that mandates are needed. The two factors partially offset.
In my previous post I recommended looking at Taiwan, as I’d hate to see libertarians put their eggs in the Sweden basket, in case public opinion there eventually forces stricter measures. The concept of no mandates is good, but requires skilled execution on other fronts. But I wasn’t arguing that Sweden’s decision not to use mandates was necessarily wrong.
In Wuhan, the government is again allowing people to go to restaurants, but they aren’t showing up. This is basically what I had in mind in several previous posts:
Wuhan’s 11 Million People Are Free to Dine Out. But They Aren’t
Is it too much to hope that the public’s excessive worry about small risks will exactly offset the externality problem of socializing during an epidemic? Of course. But that’s not the point. The real point is that we don’t know which factor is stronger, and hence we have no reason to adopt various “mandates”, at least in the current situation. By the same logic, we don’t know enough about masks to know if the government should be encouraging them or discouraging their use. So do nothing!
On the other hand, people who argue we should “take the government mandates off so the economy can flourish” are likely to be severely disappointed, as we can already see in Wuhan. President Trump won’t decide when the US re-opens. The governors won’t decide when the economy re-opens. The public will decide.
For instance, the airliners are still flying, but they are mostly empty.
What a great time for a vacation! (Just kidding.)
READER COMMENTS
David Henderson
Apr 16 2020 at 3:03pm
Excellent post. I’ll use it in my Tuesday debate with Justin Wolfers.
Scott Sumner
Apr 17 2020 at 3:22pm
Thanks David.
John Alcorn
Apr 16 2020 at 9:26pm
Re: “The real point is that we don’t know which factor is stronger, and hence we have no reason to adopt various “mandates”, at least in the current situation. By the same logic, we don’t know enough about masks to know if the government should be encouraging them or discouraging their use.”
Sweden’s Chief Epidemiologist, Anders Tegnell (video interview, 15 April 2020), discusses face masks at some length, and expresses skepticism about requiring masks outside the home, except in medical settings, elder-care facilities, and the like.
If I understand correctly, the gist of Tegnell’s argument is that a mandate for face masks outside the home willy nilly would undermine the crucial current public-health norm: ‘Stay at home if you are even slightly symptomatic!’ If there were a mask mandate, then a substantial subset of (perhaps mildly) symptomatic individuals would persuade themselves that masks adequately protects others; and they would break self-isolation and would go out with a face mask. This behavior would expose others to contagion. This anti-social behavioral effect would outweigh any modest protection from infection that masks provide.
The hour-long interview, which ranges widely, begins at cue time 00:07:30.
Cliff
Apr 20 2020 at 1:03am
“If there were a mask mandate, then a substantial subset of (perhaps mildly) symptomatic individuals would persuade themselves that masks adequately protects others; and they would break self-isolation and would go out with a face mask. This behavior would expose others to contagion. This anti-social behavioral effect would outweigh any modest protection from infection that masks provide.”
The rankest of rank speculation. Masks could very well have the exact opposite effect of reminding the wearers how serious the situation is and that they shouldn’t go out unless absolutely necessary. Also masks may be extremely effective at preventing infection of others. I just saw a story today that infection is suspected to be spread primarily by droplets and not aerosols. Masks are very effective at catching droplets.
andrew weintraub
Apr 16 2020 at 10:09pm
20 times or 10 times?
David
Apr 16 2020 at 11:59pm
20X due to OC having twice the population, but 10% of the deaths.
Richard A.
Apr 16 2020 at 10:26pm
Orange county being of lower latitude than Middlesex County certainly receives more uv radiation than Middlesex County which is probably a major cause of the different infection rate.
Here are Moscow metro employees disinfecting trains with what looks like uv lamps. https://youtu.be/ySGxmOYXcOY
Scott Sumner
Apr 17 2020 at 3:24pm
Weather plays a role, but it’s not the major cause. It might explain Orange County being 2 or 3 times lower, but not 20.
Matthew
Apr 17 2020 at 5:44pm
The climate in Massachusetts is nothing like the climate in SoCal.
If you breathe in, say, 30 virus particles and your lungs are in Anaheim weather (and hence have healthy effective mucus lining and turnover), you are likely to get antibodies and immunity but never even know you were sick.
If you breathe in, say 300 virus particles (colder, less absolute humidity means viruses last much longer in the air) and your lung mucus is dried out and not circulating in your lungs, you are far more likely to get seriously ill or die.
This is why we call it the “seasonal flu”. Weather makes an enormous difference for most airborne virus transmission and severity of illness.
John Smith
Apr 17 2020 at 2:30am
I live in Singapore, where the government releases precise data on deaths and infections on a daily basis.
Before engaging in activities that results in exposure, I will check these data and decide whether it is worthwhile to engage and what precautions is worthwhile. For example, I maybuy takeout myself with great discounts (doe to COVID driving business down) instead of delivery. I studied economics back in college.
Billy Kaubashine
Apr 17 2020 at 10:54am
The similar response to widely varied risk factors sounds like something from Kahneman’s “Thinking Fast and Slow”
Scott Sumner
Apr 17 2020 at 3:24pm
That’s right.
Alan Goldhammer
Apr 17 2020 at 2:24pm
From my reading of all the Wuhan papers (and I must have read a couple of 100 by now or at a minimum abstracts), population density and proximity tend to be major risk factors. There was an article in the Washington Post today that 25% of the beef processing industry is coming off line right now because of workers contracting SARS-CoV-2. IMO, this is something that is way underappreciated. If we did not do the shut down, there would be rolling levels of sick outs that likely would have shut down parts of the economy. I’m not a modeler and will leave it to others to figure out that economic impact. Unless we have robust testing in place, states are going to have continuing problems with the workforce. And yes, I know I am saying this for the nth time.
BC
Apr 18 2020 at 2:59am
“So even though my rational brain knows that Orange County is pretty safe and I know that 20X is a really big deal; I act on my gut, which treats the two places as equally risky.”
The death rate in Middlesex County may be 20x higher, but the survival rates are nearly indistinguishable: (1-0.000138) vs. (1-0.000007). You are 99.9869% as likely to survive in Middlesex County as in Orange County. Your gut may be more rational than you think.
Looking at ratios of death or survival rates is tricky. On the other hand percentage point differences are invariant as to whether one looks at death rate or survival rate. Also, the difference in deaths between a death rate of say 10% and 9% is far greater than between 0.0138% and 0.0007% even though 10% is only 1.11x larger than 9% and 0.0138% is 20x larger than 0.0007%. The percentage point difference in death rates, and survival rates, between the two counties is not large.
anon/portly
Apr 18 2020 at 10:32pm
If the death rate in Middlesex County is 20x the death rate in Orange Country, does that mean the relevant risk of death for a particular person is necessarily 20x higher in the former, compared to the latter?
When you think about Ro, there are spreaders, and there are spreadees. Let’s say the chance of getting infected (in say the next period in a 2-period model) is 20x higher in Orange County. Does that mean it’s 20x higher for everyone?
If, for example, each county contained two types of spreadees, one type who might get it, another type who won’t get it, for say behavorial reasons, then the probably of getting it in Middlesex would be 20x greater for the first type, but 0x greater for the second type, since the probability would be 0 for that type in both counties.
That’s an extreme and no doubt unrealistic example, of course.
Mark Bahner
Apr 20 2020 at 10:56am
Hi,
If you’re figuring risks, you need to consider whether you happen to be at a long-term care facility:
Nearly half of COVID-19 deaths in MA are at long-term care facilities
Comments are closed.