The problem of herd non-immunity

I listened to Russ Roberts and Tyler Cowen discuss the coronavirus. If you missed it, maybe you can find it at the Mercatus video archive at some point. Here is rough paraphrase (caricature?) of part of the dialogue.

Russ: Think of this as a three-week vacation. Why can’t the economy recover from a 3-week vacation?

Tyler: A lot of organizational capital will be lost.

Russ: Wha???

Tyler: Employers and employees know how to work together. When those relationships need to change, it takes a lot of time for matching and training to work out.

Russ: You seem to think that the virus will be a factor for a long time. Why?

Tyler: Suppose that in the very short run we get it under control through social distancing. That means that a lot of people will not have had it yet. What is likely to happen is that there will be a series of outbreaks, and that means a series of shutdowns. It means that it will be a long time before people feel comfortable going to locations where they will encounter crowds.

Think of this as the problem of herd non-immunity. It means that we could go a long time during which people change their behavior to avoid catching/spreading the virus.

To me, this makes two things important. One is the process of testing and approving a treatment. If a treatment can be shown to work, then we can be much more relaxed about allowing people to get the virus.

The other is having a random testing program. John Iaoannidis is getting a lot of flak, some of it deserved, for what he wrote yesterday. But I very much agree with this:

The most valuable piece of information for answering those questions would be to know the current prevalence of the infection in a random sample of a population and to repeat this exercise at regular time intervals to estimate the incidence of new infections. Sadly, that’s information we don’t have.

If the government won’t do this (and I have little confidence that they will), then I hope some corporation or non-profit will take it on.

David Henderson’s pessimistic bet

He writes,

I bet that by the end of the calendar year, the number of deaths that can clearly be attributed to the disease will be greater than 100,000.

Note that he hopes he loses his bet, as do we all.

Think of 100,000 as 10 million cases times a 1 percent death rate.

Why might we get more than 10 million cases? First, it is possible that the number of cases that have not been officially detected is 100x the number of official cases. Because some infected people do not have symptoms, some who have symptoms are not going to get tested, and some who want to be tested have been, until recently, turned away.

Even if the true number of cases in the U.S. is only 10,000 today, if that were to double ten times we would be at 10 million. If the doubling time were a week, it would take only ten weeks to be over 10 million. And as of now the doubling rate is faster than that in many European countries and in the U.S.

But given the sharp reduction in travel and large gatherings that has taken place, I expect that the doubling rate will slow down. Suppose that, after these changes have been in place for a few weeks, we find that it is taking a month or more to double the number of cases. That would make it less likely that we hit the 10 million total by the end of the year. (Although, again, it is hard to know where we are starting from.)

We might also find that we have a lower death rate. Compared with other countries, we have less smoking and more capacity in our health care system. And the steps that we take to protect at-risk populations from the virus may prove effective.

This might be a time to update the status of two hypothetical bets of mine. First, I hypothetically bet that no centrist candidate would arrive at the Democratic convention with more than 40 percent of the delegates. That now looks like a bad bet.

More recently, I hypothetically bet that the number of Covid-19 cases in the U.S. would be more than 12,000 by the end of this week. As of Tuesday evening, the total was over 6000, and it appeared to be doubling every two or three days. Unfortunately, it looks like I will turn out to be correct on that one.

Trends in poverty in the U.S.

Timothy Taylor writes,

while it might seem that evidence suggesting that that US poverty level is actually far below the official rate is good news (to the extent that it is true), nothing is simple in a politically polarized world. Conservatives would have to accept that a number of government programs have had a dramatic effect in successfully reducing poverty rates. Liberals would have to accept that poverty is now a much smaller problem than several decades ago.

I recommend the whole post for its analysis of data and concepts.

My opinions:

1. The rising tide of economic growth has tended to lift all boats, and that accounts for some reduction in poverty, as properly measured.

2. Government transfer programs, such as food stamps, have also contributed to poverty reduction. Certainly this is true numerically.

3. But government transfer programs have, in my opinion, undermined social norms regarding work and marriage. The high implicit marginal tax rates that arise as people lose eligibility for benefits when they earn income have made it uneconomical for women to marry low-wage men. So low-wage men work less and marry less than they would otherwise.

Macroeconomics of the crisis, 6

Tyler Cowen has a short paper on the topic.

The key point is that in the short run we want economic activity to fall in ways that will curb the spread of the virus. In the long run, we want economic activity to come back.

In my view, the primary channel by which a short-term reduction in economic activity leads to a long-term decline is the financial channel. I have an essay on that, which I will post below the fold. Continue reading

Macroeconomics of the virus crisis, 5

A reader writes,

I’m seeing millions of people, primarily at the middle to lower end of the economic spectrum, who will be financially ruined by the response.

While remaining humane and just, is there a more cost effective approach to the problem, particularly since it seems concentrated in a rather narrow subset of the population (those with compromised immune systems and the 60+ cohort)? E.g. instead of isolating the entire population, why not just isolate those most at risk and compensate them accordingly? How do we weigh the ethical concerns of one group as against another? Why do such questions seem so out of bounds and antisocial?

1. I think that even the “low-risk” population, if they move about freely, are likely to overwhelm the health care system. I doubt that encouraging them to go about their business is the first-best strategy. But it might work out better than other approaches.

2. What we are doing now, which is lockdown-lite, is also not the first-best strategy. It may exacerbate economic pain while failing to do enough to stop the virus. It may be an instance of my two-weeks-behind hypothesis. Still, I think we ought to support the lockdown-lite approach–as you know, my wife and I adopted self-quarantining last Thursday. We ought to give it a chance.

3. I suspect that the first-best strategy is a total, nationwide lockdown for two weeks, enforced militarily. The intent would be to deprive the virus of hosts. Even then, in order to contain subsequent outbreaks, we would have to continue to encourage social distancing, require people to keep track of contacts, and do a lot of random testing of asymptomatic people.

Think of this as comparable to the Arab oil embargo. Probably the harder the adjustment we make sooner, the better it will be. The macroeconomic policies of that era made things worse, and I expect the same to happen today.

I cringe at all the talk of monetary and fiscal “stimulus.” What this means is that governments will use the crisis to enlarge their share of the economy, which will hurt the adjustment process, not help it. It will also “solve” the problem of debt collapse by piling on more debt.

Friends who might lose benefits

From the WSJ,

More couples are deciding to live together instead of marrying, and strained finances are a top reason many cite. A survey last year by the nonpartisan Pew Research Center found that among those who live with a partner and wish to get married, more than half said they or their partner weren’t financially ready.

About half of middle earners were married in 2018, a drop of 16 percentage points since 1980. Among the highest U.S. earners, 60% were married in 2018, a decline of 4 percentage points over the same period. That marks a reversal. In 1980, a higher proportion of middle-class Americans than top earners were married.

1. You have to decide whether or not to have children.

2. You have to decide whether to live independently or together.

3. If you live together, you have to decide whether or not to get married.

It seems to me that the decision that ought to most be affected by economic circumstances is (1). Raising children is expensive. And that decision in turn would affect (2) and (3).

Whatever you decide about (1), I can also see (2) having an effect, since it is cheaper to live together. And that in turn would affect (3).

But mostly the article is written as if financial status directly affects (3). Both the headline and one of the academics quoted in the story refer to marriage as having become a “luxury good.”

I don’t see (3) as the likely margin along which financial status affects decisions. Something is wrong with this picture.

If the chain of thinking were “We’ve decided that we can’t afford children, and if we can’t afford children then there is no point in getting married,” would make sense. It also would be very sad.

But the article says:

More couples are forming families without matrimony. One in four parents living with a child is unmarried, according to Pew. More than one-third of them are living with a partner, up from one in five in 1997, the Pew study of 2017 data found.

Seriously? People are thinking We wanted children, but getting married seems like too much of a commitment. We can’t afford to make that kind of commitment yet. ?????

I still think that replacing means-tested entitlements with a UBI would make low-wage men more attractive as marriage partners. Indeed, the article profiles a couple with children who fit with my model of non-marriage.

They said they want to get married but are holding off because Ms. Dlouhy is enrolled in a publicly funded program that pays for her to earn a nursing license. Combining their income could jeopardize that assistance, she said, as well as her state health-insurance subsidies.

Herd immunity and exposure policy

Robin Hanson wrote,

it isn’t crazy to consider cutting pandemic deaths via more infection inequality, including via deliberate exposure.

Pointer from Bryan Caplan.

Consider the following strategy:

1. Separate the population into low-risk and high-risk groups, based on their conditional probability of death if they get the virus. For example, young people with healthy immune systems vs. older people and/or those with compromised immune systems. Separate them not only conceptually, but physically–don’t let anyone from one group get near someone from the other group.

2. Then, allow the low-risk group to become infected, while keeping them away from the high-risk group.

3. Once the low-risk group have recovered, let the two groups mix.

Two reasons to hesitate about doing this. One is that it is not certain that people who have had the virus are immune. There are anecdotes about people re-acquiring the disease. Perhaps there are multiple strains, rather than “the” virus.

A second reason to hesitate is the high rate of death among health care providers, many of whom are young with healthy immune systems. This suggests that there are some other factors that affect risk, and you want to know more about those other factors before you try this approach.

My own “out of the box” suggestion is a program to test a random sample of people who are asymptomatic. That would give us a better idea of the dynamics of virus spreading.

Guns and incarceration

Barry Latzer writes,

Western Europe, Canada and Australia have far fewer guns and, compared with the United States, far fewer crimes committed with firearms. From 2000 to 2012 there were an estimated 1,500 gun homicides per year in all of Europe, around 20 percent of total homicides. For a comparable period there were nearly 12,000 annual gun homicides in the United States, eight times as many as Europe, and guns were responsible for 67 percent of all killings. In 1990, a peak year for murder here, the Western European firearm homicide rate was a mere 0.53 per 100,000. The United States rate was 5.57 per 100,000—ten and one-half times higher.

He cites these statistics as part of a claim that incarceration rates in America, while high relative to that in other countries, is not high relative to the dangers that the criminal population presents.

Note that this observation will trouble just about every political view. Conservatives and libertarians will not like to see guns treated as a causal factor in high crime severity. Progressives will not like to see high crime severity as a cause of “mass incarceration.”

Clark Neily attempts to counter Latzer’s overall claim that the U.S. does not over-incarcerate, but I find Latzer’s analysis more persuasive. Let me state my views, which are tentative and impressionistic, rather than well-formed.

1. There may be some margins along which more lenient policies on imprisonment would have little adverse effect on the prevalence of crime, but I am dubious of the view that we would be better off with a much smaller prison population.

2. It is conceivable to me that at some point in the future, prison will be seen as an inhumane solution to crime. Other solutions that protect the public while giving convicted criminals better treatment will have been shown to be superior. But those “other solutions” do not yet exist, or are not widely understood of they do exist.

3. I am convinced by those who say that prosecutors play a game of threatening accused criminals with very severe sentences in order to get them to plead to lesser offenses. You should want the crime with which a prosecutor charges the defendant to be the charge with which the prosecutor expects to convict the defendant, not the charge that can scare the defendant into a plea bargain. Even if the outcomes of the current game are reasonable (i.e., the defendants who plea bargain actually deserve their sentences), the process is rotten. We should figure out what got us into this equilibrium and how to get out of it. Note that what got us into this equilibrium might include characteristics of the judicial process that make it difficult for prosecutors to succeed without playing the plea-bargaining game.

4. We have too many statutory crimes. It would be better to aim for a system with a few statutes, well and uniformly enforced. If you want to see a more tolerant criminal justice system, I would propose that you put more of your energy into the front end of the process (pruning the criminal statutes) than the back end of the process (reducing incarceration of those who are charged). For example, if you think that too many prisoners are serving over-long sentences for drug offenses, then try to change the relevant statutes.

I wish I could bet Roger Kimball

He writes,

Here’s what I think you’ll find if you track things for the next week. The number of cases will go up by 500-1000 cases per day for the next couple of days. Then the increase will start to decline. The number of deaths will also go up, but modestly. I suspect that by the end of this coming week the total number of cases will be around 6000, the total number of deaths 100-150.

I bet that the number of cases by the end of this coming week will be more than double his estimate.

Consider the following math problem:

You observe a petri dish for 24 hours. The amount of bacteria is doubling every hour. At the start, there is only a microscopic amount of bacteria. Right at the end of 24 hours, the petri dish is filled with bacteria. When do you suppose that the petri dish was half filled with bacteria?

Many of my readers will arrive at the correct answer after a few seconds. I would bet that Roger Kimball would not be able to answer the question correctly within five minutes. And if you cannot do that, you should keep your views on the virus crisis to yourself.

In his own way, Nassim Taleb tries to explain why it is unethical not to pay attention to the exponential.

In a more easily-understood way, these physicians make the point.

If our health care system fails, then we will all suffer. If the hospital is choked with COVID-19 patients, people with appendicitis, heart attacks, broken ankles, and so on will not be able to be treated. This is the picture of systemic risk. Everyone is at risk if there is a systemic failure of our health care system, not just those with COVID-19.

The challenge is this: By following the appropriate recommended social isolation measures, you will be saving lives of not just those at increased risk who are infected, but also those who need other critical health care services, including potentially yourself. You will be saving the lives of people you will never meet.

Who should follow our suggested social isolation measures? EVERYONE. If you do not need to go out for a mission-critical purpose, do not. Again, you WILL be saving the lives of at-risk members of your own family, as well as people you will never have the pleasure of meeting.

This reinforces to my thought process.

Re-reading an essay on the analogy with religion

James Lindsay and Mike Nayna wrote,

to the degree that we can accept that Social Justice is a faith-based program based upon a kind of locally legitimized special revelation, we should feel serious concerns and discomfort about institutionalizing its beliefs in any space that isn’t wholly devoted to them. We should also be quick to be honest about which spaces are and which aren’t. Public institutions like public universities, being public, should be very hesitant to implement Social Justice initiatives. Private institutions, like corporations and private universities, can make their own choices on the matter and accept the benefits and consequences of openly aligning with a faith initiative as they come.

. . .Social Justice, because it is an (applied) postmodern mythological system upon which a moral tribe is built, is not technically a religion but is a kind of faith system. This raises serious questions about how we should deal with its attempts to institutionalize itself in various cultural enterprises—especially education—under the guise of being secular in the broad sense merely because it qualifies in the narrow sense. Most importantly, however, it provides all of us with explicit permission to treat its claims and advances in the same way we would any other faith—say, like Scientology—and to proceed accordingly without the guilt it attempts to foist upon us as a conversion mechanism.

I linked to the essay when it first appeared, but when I recently came across it again I felt it deserved another mention.