What is a pandemic compared to our sewer system? An example of how a society normalizes riskspost by Ramiro · 2020-07-25T14:59:24.093Z · score: 20 (7 votes) · EA · GW · None comments
Survivorship bias Risk Aversion, Peer Pressure and the Tragedy of the Commons A Brazilian problem or an intergenerational one? (Derogatorily dubbed The Millenial Eulogy) Notes: None No comments
[This is a cross-post / translation from: https://medium.com/altru%C3%ADsmo-eficaz-brasil/o-que-e-uma-pandemia-comparada-ao-nosso-esgoto-5cbbc530b768
This might interest those who are trying to understand: a) how culture might affect development trajectories, and b) what is happening in Brazil - which has been following the lead of US in becoming a paradigmatic Covid-19 tragedy, and might even surpass the latter (seriously).
Remind this is a blog post for a Brazilian broad audience; I did edit it a little bit and took off the most obvious parts and memes in order to publish it here, though. Comments on how to improve this sort of interaction with non-EAs are very welcome. Also, forgive me if it's more personal or narrative than the usual EA Forum posts; I totally failed in avoiding it.
Epistemic status for the big statements: I'm pretty sure short-term bias is an important problem in developing countries; I have no idea on how to solve this - there's a case for the best answer being economic development [EA · GW], but I wish there was something more immediate. Finally, the association between this problem and intergenerational issues is, I hope it's clear, just a suggestion; however, I do think demographics is often underrated by development economics and related subjects]
You may recall that in March, when commenting on the Covid-19 pandemic, the President Bolsonaro stated that Brazilians can jump into sewage without harm.
I don't want to polemize it, but as I started writing this text, I just googled "death in sewage" (esgoto in Portuguese). One week before, a 3-year-old boy had drowned in a cesspit in Mafra/SC; in March, another one had died in Ariquemes/RO. Anecdote is not data, but statistics confirm this is a problem: in 2018, it is estimated that in Brazil, on average, 2.71 people died from diarrhea and 2.78 people died from drowning for every 100,000 inhabitants. Too many drowning strangers.
The problem here is not what the president said. Actually, perhaps the "butterfly effect" of his controversial statement helped unlock the new Basic Sanitation Framework project in Brazil, which was halted in Congress thanks to the political instability of the last five years. I want to point out two lines of thought (possibly) underlying this speech and which are shared by a good part of our society: survivorship bias and good-old tragedy of the commons.
In March, when the epidemic stroke Brazil, I had strong cold symptoms in the work; I picked up my stuff and told my colleagues goodbye. One of them, a man in his sixties, tried to hug me, said that I was exaggerating, and that what I had missed was a healthy childhood in the countryside, swimming among snakes in the river.
I then asked him how many of his friends had drowned; he replied, serious: "three”.
In survivorship bias, you focus only on the sample that has made it past a selection process, ignoring those who didn’t, and end up overestimating the overall probability of surviving - i.e., a more everyday version of the anthropic shadow. In the example, it wasn't so much the childhood in the countryside that made him healthy - he survived because he was lucky and robust. Another common example: when we think of companies and entrepreneurs, we focus on success stories - no one remembers that even in the U.S., 65% of companies do not survive the first 10 years.
In the 90s, I did once fall into some sort of cesspit that collected waste in my grandfather's street, in a small city. That pit no longer exists, fortunately. Hans Rosling, in his book Factfulness, also tells us how he almost drowned in a sewage pit in Sweden in the 1950s. That nightmare ended in Sweden, and it should have ended here - where basic sanitation is still a luxury. In the neighborhood my father currently lives, populated by boomers' summer houses, with bars and pools in the garden, there are still sewage pits like the one I fell into; but at least I've seen no children playing around them.
Note that, according to our Sanitation Map, though sewage network coverage evolved only from 59% to 66% of households between 2009 and 2018, hospital admission rates fell from 184.7/100k to 65.6/100k. The main bottleneck in this progress seems to be regional differences - in the state of Piauí, only 7% of households had sewage collection in 2018.
The problem with survivorship bias is that it generates unrealistic self-confidence; worse: in some societies, stressing the risks ends up being seen as paranoid or offensive. I remember the time I got into a taxi in Cusco and fastened my seat belt; the driver asked me to unfasten it: it offended him, because it signaled that I didn't trust him. Because of that comment, my confidence in him obviously diminished. However,he couldn't realize what, for most of us may sound obvious: a) if I really thought it was likely that he would crash the car, I wouldn't even have gotten in; b) almost all traffic accidents, with or without a belt, occur with people who are very sure that they will not suffer a collision.
Risk Aversion, Peer Pressure and the Tragedy of the Commons
What I am about to present now is not exactly a fallacy – quite the opposite, it is valid reasoning. Something like: if you don't have high expectations about life, if your environment is already full of risks and you don't expect to live long, then you have little to lose; it doesn't make much sense to think about your future or about uncertain events. Carpe diem.
Forgive me for this very simplified and unrealistic example: suppose someone invents a new technology (new food, sanitation revolution, etc.) that increases the average life expectancy by 5%; in a society with a life expectancy of already 80 years - then people would be willing to pay the equivalent of 4 more years of their lives. If you live in a society with a life expectancy of 40 years (i.e., more than the global average until the 20th century), this means only two years. If people value life expectancy in absolute terms (1 year more is as good in the first as in the second society), but resources in relative terms, then the first society is much more likely to pay for this innovation - because it has more to gain[i].
Something similar applies to risks of losses. Imagine that, instead of an innovation, we have a new disease that has decreased life expectancy by 5%; in the second society, you have less to lose, and so you may be less risk-averse [ii].
Let's complicate it a little bit: now, you also have more urgent things to worry about, like feeding your children. So, instead of reducing the odds of contracting this disease by isolating yourself, you will return to work; in the first week, everyone will be very cautious, wear masks, avoid contact, wash their hands 15x/day. In the second week, most people are already going to the supermarket or restaurants every day; people still wear masks, but they already form groups to talk - and you may start wondering "why not?”. In the fourth week, it is discovered that someone has been infected; you and your colleagues redouble your care, some are diagnosed, licensed... But probably, no one will think much about the other people they met at the supermarket and restaurant.
I'll highlight two points from this second scenario:
a) peer effect: your perception of risk has changed as you have observed the behavior of your colleagues change, and vice-versa - i.e., we tend to act like the people we perceive as resembling us.
b) Tragedy of the commons: your individual behavior (your unilateral decisions) started having less of a relative impact on your own risk exposure - after all, if your colleagues were already exposing you to a higher risk, it doesn't make such a difference, marginally, if you are avoiding going to the restaurant. Ultimately, there is a feedback loop between the level of risk people are already exposed to, and the additional risk they’re willing to accept.
I want to highlight this is a more profound problem than it seems at first sight: prosperous societies have better risk management and greater concern for the future; they prepare for natural disasters that occur only once in a century, create sovereign funds to finance future generations, have lower interest rates, more long-term investment, greater social confidence, less instability... which makes their future more prosperous, and so on. These are societies where people no longer fall into sewage pits.
For most of history, human societies were not like this; they were societies where (to use Keynesian lingo) the marginal propensity to consume additional income is close to 100%. In R. Shiller's description (Narrative Economics, Cap. 9):
Most people then had no concept of retirement or sending their children to college, so they had no motivation to save toward these goals. […] If they became bedridden in old age, they expected to be cared for by family or by a local church or charity. Life expectancy was short, and medical care was not expensive. […] The people saw the authorities as responsible for instilling moral virtues rather than building consumer confidence.
That reminded me Pygmalyon / My fair Lady, where Eliza Doolitle's wise father, when trying to extort Henry Higgins, warns that he prefers a low amount he can spend quickly - for if he received too much he would start thinking about how to save it, and so be forced to change his life habits. In fact, at the end of the story, he curses Higgins for having arranged for him to receive a small fortune, so guaranteeing his material security - but also making him a married bourgeois, concerned with his own reputation and savings. Shaw's character is more ironic than derogatory: despite being a rascal, Mr. Doolitle is very clever and articulate, and he becomes a gentleman (though he deplores it) as he becomes rich; actually, the whole play is about how the environment shapes behavior in very subtle ways.
A Brazilian problem or an intergenerational one?
(Derogatorily dubbed The Millenial Eulogy)
It's quite visible that voting patterns (and likely current political polarization) have an intergenerational component - youngsters are less likely to be conservative. This can be shown by the support of the government according to age:
I feel I often fail to show due respect to boomers; they came from another world, where falling into sewage was “just another Tuesday.” A world still haunted by terrors such as smallpox, war, famine and massacre. I wonder how much this has psychologically affected them: poverty and scarcity impact the way people think throughout their lives. Life expectancy in Brazil increased from 54 to 75 years between 1960 and 2019; boomers know the world where they grew up is over, but they may still fail to adjust their expectations and risk-attitudes in accordance with this change. Just like what passes as avarice is often the neglect of inflation, they often fail to account for the fact people are expected to live twenty years longer[iii]. Just yesterday, my mother was complaining about how tragic was the world: her aunt's, such a kind person, had lost her firstborn (of seven kids) to (probably) polio; then I made Mom realize, for the first time in her life, that no one has died from polio in Brazil in the last 30 years.
My impression is that young suffer from a distinct bias, which sometimes leads to a similar effect; think about how impatient and reckless children are because they have not yet internalized the fact that they have much time ahead of them. They form their expectations from the little they have lived - they model life on their past: for someone ten years old, one year is seen as 10% of their life, not 1.4% of what is still ahead. Of course, lack of experience also prevents them from having more accurate notions of risks: they have already been told that fire burns, but they have never had this experience. No wonder accidents have a significant impact on the mortality of children and young people (up to 30 years old).
That's why, from a more collective point of view, I regret that our educational system sometimes miserably fails to convey valuable lessons from History. I remember my classes never mentioned the scale of the 20th century progress, nor the horrors left behind, like polio and smallpox, but en passant - perhaps because we needed to learn, for the test, the name of some mediocre politician or a general who had won an irrelevant battle. I thought disasters were just nightmares from which we had woken up, that could no longer affect us from its inexistence; and I was one of the students who loved History.
It takes even more time for generations to learn from history. In 1910, Norman Angell [iv] published a best-seller, The Great Illusion, arguing that wars were a thing of the past: the Pax Mercatoria (stable international relations provided by free trade) had made war something irrational, for even a victorious aggressor nation would face bitter economic and welfare loss - something analogous to the current notion of mutually assured economic destruction.
Many have interpreted this "irrational" as "impossible"; and the Gilded Age, the Belle Époque, has remained in memory as a Golden Age of prosperity and peace[v] ... and the rest of the century was in fact marked by an unprecedented, almost utopian improvement in life prospects - especially thanks to the Health Revolution, which reduced the incidence of infectious diseases and the likelihood of someone falling into the sewer. It is still underway in Brazil.
But four years after Angell's book was released, the lost generation plunged into World War I, the Tragedy of the Century, with some 22 million dead; and this helped to spread the H1N1 Influenza around the world, which killed another 50 million people, at least. It should have been the "war to end all wars", but it ended up leading to World War II, a generation later - which in turn brought totalitarian systems and the Cold War into the world.
We should learn from the experiences of previous generations - from what really happened, not just what they like to remember. The main lesson I draw from the 20th century is that these are real threats, but they can be overcome - if we are careful, if we make flexible plans to defeat them. So the future may continue to be better than the past. Sometimes these topics sound a little gloomy... but I can't help but feel ridiculously happy when I see the progress made in this century, and we weren't even really trying.
One day my children will think that drowning in sewage is as impossible as dying of smallpox; and I will teach them that this is only true if they are willing to continue the progress that has been made (or perhaps to destroy the smallpox samples in laboratories with limited security).
Most profound thanks for revision and proofreading: Fernando Moreno, Gabriel Espíndola and Angela M. A. B.
Disclaimer: only the author is responsible for eventual mistakes or hazards related to this text, and under no circumstances can responsibility be attributed to other people.
[i] This is particularly important for the definition of metrics to assess the impact of health interventions (or even welfare in general), where the "value" of lives saved per monetary unit invested must be compared. Currently, the two most fashionable metrics are DALY (disability-adjusted life years) and QALY (quality-adjusted life years), which consider the increase in the expectation of lives saved by some model that considers deficiencies, diseases or situations that impact the quality of life. The tragic implication is that the value that is attributed to saving the life of a certain person can be correlated to the place where that person lives (besides the existence of previous diseases, age, etc.). It is common that, when we say "an intervention saves 1 life for every $3,000.00", by "life" we mean something like "30 QALYS".
[ii]Warning: this is obviously a simplification. Also, this is not an allusion to the Covid-19 pandemic per se; I only use this example because, today, it is easier to imagine.
[iii] This can be associated with a certain neglect of inflationary effects. My mother recently complained about wine prices, and it became clear that the problem was that her concept of "cheap / expensive wine" was formed 20 years ago, when a R$100 wine would be something absurdly elitist.
[iv] Angell himself won the Nobel Peace Prize in 1933 - the same year Hitler came to power. Later, like other pacifists, Angell adopted a more realistic position and advocated for a direct confrontation with Germany in World War II.
[v] For those who were not under colonial rule, of course.
Comments sorted by top scores.