Book Review: Lost Connections

The actual content of Lost Connections: Why You’re Depressed and How to Find Hope by Johann Hari is significantly less self-help-y than the title would suggest. If I were to summarise my main takeaway from this book, it would be this: people are mostly depressed because their lives are bad. Lost Connections is about how antidepressants are wildly overprescribed, and how Big Pharma has marketed them as a panacea using dodgy science while ignoring the complex social and economic roots of depression and anxiety. There are parts of the book that I liked, but I have some problems with it.

What is depression, anyway?

One of the things people say about depression is that it’s a “chemical imbalance” – usually, a lack of serotonin. Needless to say, this is nonsense. It’s unclear what it would even mean for the brain to be in a state of “chemical imbalance”. Also, while serotonin is known to have something to do with depression, it’s not a straightforward relationship: if you give a chemical cocktail to normal people which lowers their serotonin, they don’t get depressed. Also, tianeptine is a common antidepressant in Europe which works by lowering your serotonin. Some claim that psychiatrists used to believe in the “chemical imbalance” theory but have since moved on. This seems like a strawman and I can’t find evidence of it ever being widely believed.

So that’s one confusion about depression. Another is about the extent to which depression is psychological vs. physical. Of course, the brain is physical, and its behaviour is completely determined by the laws of physics. So, in a trivial sense everything is equally physical. What it means to call something ‘psychological’ is actually quite philosophically complex. It means something like this: there are multiple emergent levels of reality. For instance, atoms are real, and presumably chairs are real too. Something can be considered psychological insofar as it’s more parsimonious to consider it with respect to the mind/consciousness level of reality, rather than the cells/biology level of reality.

There are two extreme ways of looking at depression. One is that it is caused purely by one’s thoughts in a way that is entirely divorced from the physical world – this is the naive view Hari attributes to most doctors in the past. Another is that depression is “just” like a physical illness and should be treated as such. I hear this kind of language from people sometimes, and one of the motivations for saying this, I gather, is that if people think of mental illnesses in much the same way they do, say, cancer, then there wouldn’t be such a stigma surrounding it. But Hari points to research showing that things actually become more stigmatised when they are thought to result from unchangeable biological characteristics rather than development. I discussed this research (Mehta 1997) with friends recently, and the best explanation anyone could come up with was that humans are tribal, and people who are genetically or biologically different to us are an out-group, whereas those who had bad things happen to them or fall into a negative cycle may well be in our in-group. The next step would be to check whether congenital conditions get less funding relative to the proportion of the disease burden they represent.

Ethan Watters discusses stigma in his excellent NYT piece The Americanization of Mental Illness. He points out that, while the social acceptance of many forms of mental illness has grown, for others acceptance has actually fallen:

“At the same time that Western mental-health professionals have been convincing the world to think and talk about mental illnesses in biomedical terms, we have been simultaneously losing the war against stigma at home and abroad. Studies of attitudes in the United States from 1950 to 1996 have shown that the perception of dangerousness surrounding people with schizophrenia has steadily increased over this time. Similarly, a study in Germany found that the public’s desire to maintain distance from those with a diagnosis of schizophrenia increased from 1990 to 2001. Researchers hoping to learn what was causing this rise in stigma found the same surprising connection that Mehta discovered in her lab. It turns out that those who adopted biomedical/genetic beliefs about mental disorders were the same people who wanted less contact with the mentally ill and thought of them as more dangerous and unpredictable. This unfortunate relationship has popped up in numerous studies around the world. [A] study, which looked at populations in Germany, Russia and Mongolia, found that “irrespective of place . . . endorsing biological factors as the cause of schizophrenia was associated with a greater desire for social distance.””

On the flip side, there is the worry of concept creep: when people have a term for something, the set of phenomena that it refers to tends to grow over time. So, what appears to be a mental health crisis could just be a broader class of symptoms being regarded as mental illnesses. Concept creep is, in a way, what happens when there isn’t enough stigma – at least, enough stigma surrounding carelessly diagnosing oneself or others with a mental illness.

So, there are downsides to couching depression in biomedical language, but it is of course to a large extent biological. We know from adoptive twin studies, for instance, that depression and anxiety are 30-40% genetically determined. When I mention the “biological” causes of depression, you probably think of sleep, diet and exercise. But there are other more obscure factors: people sometimes get depressed as a side-effect of medications to treat unrelated conditions, or because they’ve been exposed to lead, and their depression sometimes goes away after they start using really bright lightbulbs.

Antidepressants (mostly) don’t work

Anecdotally, for most people who take antidepressants, it’s hard to tell whether or not they’re working. But for some subset, it helps substantially and is sometimes utterly life-changing. Given this, you would expect that, in clinical trials, you would see a moderate effect size from antidepressants. But nope: you see a tiny one that is frequently indistinguishable from placebo. (Side note: effect size is frequently mistakenly thought to mean “the size of the effect”. But effect size in statistics is this number divided by the standard deviation, making it a dimensionless quantity.) How do we square this with these anecdotal reports? Part of the answer lies in the variance of outcomes. The rule of thumb seems to be that one third of people have undeniable positive effects from their first antidepressant and two thirds of people eventually get the same from one of the antidepressants that they try. These studies are averaging across an entire group, hence why we see such a small effect size. There are also more technical points about study design – for instance, if the study is not of first-time takers, then those individuals with particularly intractable depression will be overrepresented among those being studied. Another part of the answer is regression to the mean: people are likely to seek out medical help when they are at a low points of their lives, and so things will likely get better due to pure chance, which patients (and their doctors) may well think is an actual effect of the drug. When people talk about studies containing a placebo group, they really mean two things: some mysterious psychological force whereby the expectation of something causes it to happen, and regression to the mean. There is a fascinating body of evidence showing that, often, this get-better-anyway effect is much larger than the bona fide placebo.

So, the evidence for the efficacy of antidepressants is quite weak, but we also have these reasons to believe that they’re somewhat helpful. Plus, even if it’s mostly a placebo, that’s still worth it as long as people feel better, right? Well, Hari says, this might be correct if it weren’t for the very real side-effects of antidepressants. There are the usual side-effects you might get from any drug, like nausea and fatigue, but notably also sexual dysfunction. Sexual side-effects are particularly common among takers of selective serotonin reuptake inhibitors (SSRIs), which is the most common type of antidepressant and what most people think of when they think of antidepressants. The drug companies who produce these medications have a pretty strong incentive for exaggerating their effect and downplaying the severity of the side-effects. Also, as we’ll see, the side-effects of long-term antidepressant use are not well understood.

The number of people who take drugs for psychiatric problems is pretty shocking: 20% of US adults are taking a psychiatric drug, 25% of middle-aged women are taking an antidepressant, and 10% of boys in high school are using prescribed stimulants to help them focus. These numbers are so high that I don’t believe them. The US government says that antidepressant usage is 10%, although the figure for middle-aged women is, alarmingly, basically accurate.

Hari introduces us to the foremost critic of antidepressants, Irving Kirsch, author of The Emperor’s New Drugs. Hari summarises his research as concluding that the effects of antidepressants are 50% placebo, 25% regression to the mean, and 25% real effect. His arch-nemesis, Peter Kramer, is antidepressants’ foremost defender:

“[Peter Kramer’s] first argument is that Irving is not giving antidepressants enough time. The clinical trials he has analyzed—almost all the ones submitted to the regulators—typically last for four to eight weeks. But that isn’t enough. It takes longer for these drugs to have a real effect. This seemed to me to be an important objection. Irving thought so, too. So he looked to see if there were any drug trials that had lasted longer, to find their results. It turns out there were two—and in the first, the placebo did the same as the drug, and in the second, the placebo did better.”

I think it’s appropriate to be flabbergasted that there are (or at least, were) only two studies that lasted for more than eight weeks of a type of drug that 10% of Americans use. Even Kramer doesn’t agree with the current regimen of keeping people on antidepressants for basically their whole lives (for context, Hari has a history of depression and had been taking antidepressants every day since he was a teenager):  

“Even Peter Kramer had one note of caution to offer about these drugs. He stressed to me that the evidence he has seen only makes the case for prescribing antidepressants for six to twenty weeks.”

Next, Hari talks about 5-HTT, a gene that was thought by the research community to be significant in explaining depression but which, according to this Slate Star Codex piece, turned out to have no effect whatsoever. (This article is about 5-HTTPLR, and I can’t figure out if this is the same thing as 5-HTT. I know very little about biology and I can’t understand the first few links on google.) Hari says that 5-HTT, and other genes that are risk factors for depression, work by multiplying the risk of depression in response to negative life events, raising the probability of a depressive episode following a major negative life event from (say) 15% to 20%. I would be interested to see how people’s genes cause them to shape their environment differently, thus making negative life events more or less common.

I mentioned earlier the extreme view that depression is entirely unrelated to the physical world. He writes:

“Michael [Marmot, the Australian psychiatrist] would walk around the hospital wards and think—all this sickness and distress must tell us something about our society, and what we’re doing wrong. He tried to discuss this with the other doctors, explaining that he believed that with a woman like this patient, we “should be paying attention to the causes of her depression.” The doctors were incredulous. They told him he was talking rubbish. It’s not possible for psychological distress to cause physical illnesses, they explained. This was the belief of most medical practitioners across the world at that time.”

Marmot went on to conduct a study that looked at UK civil servants in Whitehall. Anyone can notice that the poor and those with difficult and unpleasant jobs tend to have worse mental and physical health. But this could be for lots of reasons. All of the civil servants studied had somewhat similar lives, pay on the same order-of-magnitude, but massive differences in status, and the extent to which they had control over their jobs:

“After years of intensive interviewing, Michael and the team added up the results. It turned out the people at the top of the civil service were four times less likely to have a heart attack than the people at the bottom of the Whitehall ladder . . . If you worked in the civil service and you had a higher degree of control over your work, you were a lot less likely to become depressed or develop severe emotional distress than people working at the same pay level, with the same status, in the same office, as people with a lower degree of control over their work.”

There’s no problem so bad overregulation can’t make it worse

Hari never accepts the conclusion that a lot of his evidence appears to be pointing to: that the government, and other regulatory bodies like institutional review boards, have been slowing progress in mental healthcare for decades. The two most exciting recent developments in the fight against depression are the use of ketamine and psychedelics. Psychedelics were schedule 1’d by the US government, which dried up the research funding for decades. Ketamine is also illegal and extremely difficult to get by prescription, despite its miraculous ability to treat (via injection) intractable forms of depression.

The regulatory environment seems to be in a worst-of-all-possible worlds situation, where the fact that drug companies are really desperate to show that their drug works results in byzantine regulation to stop them from exploiting people or defrauding anyone, but the government itself won’t cough up the money to just test what actually works. He writes:

“When [a drug company] wants to conduct trials into antidepressants, they have two headaches. They have to recruit volunteers who will swallow potentially dangerous pills over a sustained period of time, but they are restricted by law to paying only small amounts: between $40 and $75. At the same time, they have to find people who have very specific mental health disorders—for example, if you are doing a trial for depression, they have to have only depression and no other complicating factors.”

Hari points out that there are basically zero large clinical trials which test the major antidepressants against one another, through a weird sort of market failure where no-one has an incentive to do this, and even if the government or a philanthropist wanted to do this, the regulations are such a pain in the ass that they don’t.

If antidepressants were mediocre but there was no alternative, then their frequent usage would be no mystery. The problem is that we know other solutions are more effective. For instance, one of the common scales for depressive symptoms is the Hamilton scale, which runs from 0 (perfect bliss) to 51 (perfect misery). Antidepressants, on average, produce a 1.8-point jump on the Hamilton scale, while having a regular sleep schedule produces a 6-point jump. Hence there needs to be some reason why antidepressants are used in excess of how useful they actually are.

Hari’s answer is corruption combined with people looking for easy answers. Indeed, almost everyone involved has bad incentives: he points out how 40% of regulators’ wages are paid by drug companies in the US, and the figure is 100% in the UK (!). I’m not sure whether he would agree with this, but it seems like he’s hinting that regulatory agencies are too liberal when it comes to approving new drugs. But I’m the kind of person that reads many blog posts arguing that the FDA (and by extension the EMA) is too conservative. Indeed, a priori, it would be surprising if drugs were approved too quickly on average. Regulators face much harsher consequences for pursuing a policy that actively leads to harm rather than by making an omission that leads to people being harmed.

Why are we getting more depressed?

Johann Hari takes it as a given that people these days are more depressed than they used to be. I’m not so sure; suicide is declining almost everywhere, in some places massively so. The past seemed pretty crap. And yes, self-report studies show mixed results, but self-report studies are almost worthless. Nonetheless, insofar as his premise actually is true, he offers multiple possible explanations. The first is that we’re more materialistic:

“[A] social scientist named Jean Twenge . . . tracked the percentage of total U.S. national wealth that’s spent on advertising, from 1976 to 2003—and he discovered that the more money is spent on ads, the more materialistic teenagers become.”

I would hope that there were corrections done in this study to try to show causality. Hari isn’t good at summarising the results of studies, and he frequently uses vague language. Regardless, this seems somewhat plausible. Advertising is, in a way, a business model based on making you feel insufficient. Being materialistic wouldn’t make people unhappy by itself, but the theory would be that it causes people to chase extrinsic goals like wealth and not intrinsic ones like fulfilling relationships, and therefore they don’t get any happier:

“The results, when [the psychologist Tim Kasser] calculated them out, were quite startling. People who achieved their extrinsic goals didn’t experience any increase in day-to-day happiness—none. They spent a huge amount of energy chasing these goals, but when they fulfilled them, they felt the same as they had at the start. Your promotion? Your fancy car? The new iPhone? The expensive necklace? They won’t improve your happiness even one inch. But people who achieved their intrinsic goals did become significantly happier, and less depressed and anxious . . . Twenty-two different studies have, in the years since, found that the more materialistic and extrinsically motivated you become, the more depressed you will be.”

A second explanation is that we’re lonelier:

What [John Cacioppo] wanted to know was—would isolated people get sicker than connected people? It turned out that they were three times more likely to catch the cold than people who had lots of close connections to other people . . . What John’s experiment found was later regarded as a key turning point in the field. The people who had been triggered to feel lonely became radically more depressed, and the people who had been triggered to feel connected became radically less depressed . . . It turned out that—for the initial five years of data that have been studied so far—in most cases, loneliness preceded depressive symptoms.”

The evidence appears to be pretty good that loneliness does in fact cause depression, rather than people getting depressed for some other reason and consequently withdrawing from society and becoming lonelier. The increase in loneliness and decline in social capital, most particularly in America, has been well-documented, most famously in Robert Putnam’s Bowling Alone. The amount of time people spent with their families has also dropped, and these trends are true in most of the developed world. This is not necessarily evidence of increased loneliness, because loneliness is not the same thing as being alone: indeed, Hari says that the correlation between how many people you know and speak with, and how lonely you feel, is actually quite weak.

It’s obvious that Hari is quite left-wing, and he identifies much of this decline in social capital with the excesses of post-1970s neoliberalism. The mechanism for this isn’t made clear – maybe the growth of an individualist mindset that is more focused on individual consumer experience than on social experiences like family and church? He has some obligatory digs at Margaret Thatcher and he approvingly gives an example of people improving their mental health via community organising… to lobby for rent control.

Depression and grief

If your mother dies, we might say it’s “justified” for you to feel depressed for a while, but if your life is going fine but for some reason you feel terrible all the time, that’s “unjustified” and therefore should be treated. So, where do we draw the line between a normal reaction to tragic things happening in your life, and bona fide mental illness?

“After you lose (say) a baby, or a sister, or a mother, you can show these symptoms for a year before you are classed as mentally ill. But if you continued to be profoundly distressed after this deadline, you will still be classified as having a mental disorder. As the years passed and different versions of the DSM were published, the time limit changed: it was slashed to three months, one month, and eventually just two weeks.”

He then goes on to mention how, in the DSM-V, the latest version, this proviso has been eliminated and you can be diagnosed with depression irrespective of your life circumstances. Hari hints that this is because the people who write the DSM are robots who don’t understand that humans sometimes feel negative emotions in response to bad events. But the DSM makes a deliberate (albeit, controversial) decision to prescribe entirely on the basis of symptoms and not on the basis of aetiology. The benefit of this is that we can just list what the symptoms are of certain mental illnesses and what has helped to treat them, rather than allow psychologists to become arbiters of what counts as a “reasonable” or “proportionate” emotional response to different life events. Nevertheless, there’s a good point here, which is that psychologists and psychiatrists have historically not given sufficient attention to how people’s life problems arise from their circumstances, diet, exercise, sleep, and so on. Hari’s own experience of the medical system seems to be particularly bad in this respect:

“As [the researcher Joanne Cacciatore] said this, I told her that in thirteen years of being handed ever higher doses of antidepressants, no doctor ever asked me if there was any reason why I might be feeling so distressed. She told me I’m not unusual—and it’s a disaster.”

The solution to this dilemma – that many depressed patients were having perfectly understandable reactions to negative life events – was to divide depression into “reactive depression” (in response in life events) and “endogenous depression” (that comes on for seemingly no reason). Needless to say, this dichotomy has been quite problematic, mostly because the things someone is reacting to in becoming depressed can be quite subtle:

George [Brown] and Tirril [Harris] explained that they had, all along, been studying women who had been classified by psychiatrists as having “reactive depression” and women classified as having “endogenous depression.” And what they found—when they compared the evidence—is there was no difference between them. Both groups had things going wrong in their lives at the same rate. This distinction, they concluded, was meaningless.”

So, is endogenous depression just fake news? Maybe. To be fair, Hari talked to a number of people about endogenous depression, and they gave a range of answers, ranging from thinking the distinction is meaningless to thinking that endogenous depression is real but makes up a small subset of depressives.  


Hari is great at pointing out the extent to which we do not currently have a pill that you can take that will make you magically happier. But I think he fails to appreciate how amazing it would be if we did, and how this should be a top priority for science. A quote from Joanna Cacciatore sums up his position pretty well:

“Our approach today is, Joanne said, “like putting a Band-Aid on an amputated limb. [When] you have a person with extreme human distress, [we need to] stop treating the symptoms. The symptoms are a messenger of a deeper problem. Let’s get to the deeper problem.””

I get this at an individual level. People want easy answers. They don’t want to be told that they’ve made many bad life choices that will be difficult to undo, or that they need to lose weight or get better friends. Or worse yet, that their woes are a necessary consequence of free trade and capitalism. Maybe this is just because Hari and I inhabit different worlds, but if anything, at a societal level this seems like the opposite of the attitude that we take. Most people are far too quick to jump to the conclusion that there must be something wrong with society, and vehemently try to avoid the possibility that there is something wrong with them. On my old blog, I wrote a piece that argued that prestigious universities should use a partial lottery to allocate places. I’m not sure whether I particularly endorse this, but nonetheless, the most common response was that this would only be a bandage on the true problem, and that to really fix this, we would have to invest in education and eliminate the discrepancies that led to the rich and privileged being so overrepresented in elite universities. From my perspective, it seems like people are absolutely desperate to go on multi-decade long questionable social engineering projects, and they don’t want to put bandages on problems enough.

The old way of thinking was to blame depression on personal failure. The new way of thinking is to blame it on nature. Hari wants us to blame nature less and society more, and he doesn’t say much about the personal failure part. I’m sceptical of efforts to blame it on any of these, and I think it’s more like we’ve been bequeathed with a tragic mismatch between all three.

Antidepressants are almost certainly overprescribed, and this is almost certainly because doctors have bad incentives. If your doctor were actually incentivised to make you healthy, healthcare would look very different. They’re incentivised to make you relatively healthier while minimising risk of malpractice lawsuits and not offending you or your parents too much. If a parent comes in describing how their teenage son is feeling depressed, the correct response may well be to point out how it would be a miracle to have such an annoying mother and not be depressed. But given the incentives the doctor faces, the correct response is to just shut up and prescribe him Zoloft. Maybe I should get in while the market is young and start selling t-shirts that say “Shut up and prescribe Zoloft”.

Thanks for Gytis for reviewing a draft of this piece.

A Beginner’s Guide to Miles Davis

Inspired by: A beginner’s guide to modern art jazz

Miles Davis (1926-1991) was a jazz trumpeter, bandleader, and composer. He was one of the most influential figures in the history of jazz, and he had a prolific output (just on Spotify, he has over 1,000 songs). Many of these are different recordings of the same song, but jazz is so improvisational that it’s difficult to draw the line of what counts as a distinct song.

My aim with this guide is to write something that would have been very helpful to me when I started listening to Miles Davis. I will list and discuss the albums that I view as essential listening, and bullets beneath will list my favourite songs from that album. This list is far from comprehensive and there are many albums I chose to leave out. This means that the jazz fusion period is underrepresented, because I personally don’t like it as much as his earlier work. Miles was famously difficult, rude, and beat his wives on multiple occasions. So do not take this piece as an endorsement of him as a person. If you feel that I missed something important, please don’t hesitate to contact me.

This post is grouped chronologically within periods, but the different periods overlap so it’s not strictly chronological. Some of the years in this post may be confusing because of the lag between recording and release. The years I mention are from the date of release. This difference is most apparent for Cookin’, Relaxin’, Workin’ and Steamin’ with the Miles Davis Quintet, all of which were recorded over a two-day session in 1956 (!) to complete Davis’ contract with Prestige records.

There is a perception that jazz is dead, and Davis himself even famously declared that jazz was dead. This is pretty unfair. For one thing, almost all of jazz is on Spotify now at amazingly high quality. For another, the decline of the cultural centrality to jazz has led to a decline in the price of concert tickets, etc., such that there’s high returns to having expertise. Tyler Cowen writes that “current times are the very best for jazz, ever”.

Jazz is so interesting to me because of its fusion of intricate underlying structure with improvisation and spontaneity. As Ken Burns put it, jazz is “familiar, but brand new every night”. Moreover, I enjoy the intellectual demandingness of jazz as a genre. Jazz musicians seem to be the most thoughtful and intelligent of any genre. Many of the more Avant Garde songs mentioned in this post don’t sound good unless you’re really concentrating. Some of it sounds cacophonous to a newcomer. This is why jazz is considerably more difficult to get into than other genres and has a lack of listenership among the youth.

Disclaimer: I’m a philistine with limited musical knowledge or ability. This guide is by no means meant to be authoritative. But I worked very hard on it. So, after hundreds of hours of listening, I present to you: a beginner’s guide to Miles Davis. You may find this guide significantly more helpful if you follow along with this playlist on Spotify, which compiles all of the featured songs in order.

The Early Years (recorded pre-1955)

Miles Davis’ career spanned the most important eras in jazz. He replaced Dizzy Gillespie as trumpeter for the quintet of the legendary saxophonist Charlie Parker, which is how he first came to notoriety. There’s a lot of static on the earlier recordings. This doesn’t bother me so much anymore, but in any case, I recommend remastered versions where you can find them. My overall highlights from this era are Boplicity, Four, and Bags’ Groove.

The Musings of Miles (1955)

  • Will You Still Be Mine
  • Green Haze
  • A Night in Tunisia

Blue Haze (1956)

  • Blue Haze
  • Four
  • That Old Devil Moon

Collectors’ Items (1956)

  • In Your Own Sweet Way
  • Compulsion

In general, there are many remastered, extended and reissued versions of all of Davis’ popular songs and albums, and frequently no canonical version. Luke Muehlhauser talked about this on his blog, which I recommend.

Bags’ Groove (1957)

  • Bags’ Groove
  • But Not for Me
  • Doxy

If you listen to this album, it’ll probably be the Rudy Van Gelder (RVG) remaster. Van Gelder was a legendary audio engineer known for editing in such a way that produced a distinctive sound.

Birth of The Cool (1957)

  • Boplicity
  • Move
  • Israel
  • Moon Dreams

As far as I know, this album played an important part in the etymology of the word cool. Davis, with his suits, beautiful women, and suave look, was the definition of cool. He was part of what it meant to be cool. In his later more experimental years his clothing was much more unusual, colourful, and counter-cultural. Another interesting thing about this album is how Davis exemplified a kind of 50s masculinity, but his music was disarming and romantic. It was very common for couples to go to see him in jazz clubs together.

Miles Davis and the Modern Jazz Giants (1959)

  • ‘Round Midnight
  • The Man I Love
  • Bemsha Swing

The First Great Quintet (1955-59)

Miles Davis’ band repeatedly shifted in its composition, but it can be roughly grouped into two stable periods: the First Great Quintet and the Second Great Quintet. There is a famous debate among Miles Davis fans over which quintet is better. The First Quintet had Red Garland on piano, Paul Chambers on bass, Philly Joe Jones on drums, and John Coltrane on tenor sax. Cannonball Adderley later joined with alto saxophone (making it a sextet). Davis had a reputation for featuring up-and-coming unknown artists on his album, and he launched Coltrane’s career. If you are only vaguely familiar with Davis sound, it’s likely this is the period that you recognise. It includes Kind of Blue, the best-selling jazz album of all it. Bill Evans replaced Red Garland on piano for some of this period, most famously on Kind of Blue. My overall highlights from this era are ‘Round Midnight, Walkin’, Milestones, My Funny Valentine and So What.

‘Round About Midnight (1957)

  • ‘Round Midnight
  • Bye Bye Blackbird
  • Ah-Leu-Cha
  • All of You
  • Dear Old Stockholm

‘Round Midnight by Thelonious Monk is one of the most famous jazz standards. Miles was the origin of many jazz standards, including Milestones and So What.

Walkin’ (1957)

  • Walkin’
  • Solar
  • You Don’t Know What Love Is

In general, title tracks really are better on average.

Cookin’ with the Miles Davis Quintet (1957)

  • My Funny Valentine
  • Airegin
  • Blues by Five

Milestones (1958)

  • Milestones
  • Straight, No Chaser
  • Two Bass Hit
  • Billy Boy

One interesting feature of jazz is that it’s a fundamentally American genre. American songs dominate popular music, and especially so with jazz. Some of this is because of the specific role played by race relations in jazz history. But one could speculate that there’s a deeper reason. America, at its best, has separation of powers and constitutional protections (or, rather, it has a longer history of this than other developed countries). It’s all about error-correction and human fallibility. Jazz, likewise, is all about revising and improvising. Contrast this with continental Europe, which has spent much of its history falling prey to one utopian ideology or another, with classical as the dominant music of high culture.

Relaxin’ with the Miles Davis Quintet (1958)

  • It Could Happen to You
  • If I Were a Bell
  • I Could Write a Book

It’s worth mentioning that the mid-to-late 50s was the apex of Davis’ cool jazz period. This is much lower tempo than bebop, which is characterised by fast continuous saxophone melodies and is where Davis got his start. Cool jazz is what people might think of as ‘coffee table jazz’. I sometimes work while listening to it, but other sub-genres within jazz are too fast-paced and complex for me to listen to while concentrating on something else. Some jazz purists would disdain the idea of listening to jazz while working at all, as opposed to giving it your full attention. Indeed, before maybe a year ago, I had never intensely listened to music while doing nothing else for any significant period, because I got bored too easily. But now I usually listen to jazz while browsing the album covers, and very often with my eyes closed.

Kind of Blue (1959)

  • So What
  • Freddie Freeloader
  • Blue in Green
  • All Blues
  • Flamenco Sketches
  • On Green Dolphin Street
  • Fran-Dance
  • Stella By Starlight
  • Love for Sale
  • So What – Live at Kurhaus

This was a major force in the introduction of modal jazz, characterised by switching among musical modes. It’s essential to listen to all of it. The last few songs listed here are from the extended edition of the album. I recommend the extended versions of most of these albums, where one exists, and listening to alternate takes of the same song. Sometimes the albums also include banter from the band, in which you can hear Miles’ famously raspy voice, which he acquired because of a throat condition.

Workin’ with the Miles Davis Quintet (1960)

  • It Never Entered My Mind
  • In Your Own Sweet Way
  • Trane’s Blues

If you really liked any of the music from before this point, you’ll probably like most or almost all of it. One of the great things about jazz is that there’s a functionally infinite amount of top-tier jazz, supposing you don’t have extremely niche tastes, while with almost everything else I consume I struggle to find content I love that I haven’t already read/watched/listed to.

Steamin’ with the Miles Davis Quintet (1961)

  • When I Fall in Love
  • Salt Peanuts
  • Surrey with the Fringe On Top

Salt Peanuts has been stuck in my head a lot recently. Because jazz usually doesn’t have any vocals, it’s harder at first for songs to get stuck in your head or to tell them apart. With time, I’ve gotten a lot better at this. There are certainly excellent jazz vocalists – Ella Fitzgerald, Billie Holiday, Chet Baker. But there is something to be said for consciously choosing to not have any lyrics in your music. Asking why jazz doesn’t have any words strikes me as a bit like asking why novels don’t have any pictures. The music speaks for itself.

Someday My Prince Will Come (1961)

  • Someday My Prince Will Come (yes, it’s based on the theme from Snow White)
  • I Thought About You
  • Old Folks
  • Teo

This may be my favourite album cover of all time. In general, the album covers from the golden era of jazz are absolutely gorgeous. This seems like a lost art because album covers are so much less prominent in the digital era. The Sketches of Spain, Miles Ahead, and Birth of the Cool covers are all iconic. I think that one of the first things that drew me to jazz, before I had any appreciation for the music, was the way people looked when they were playing it. They just looked so cool!

Collaboration with Gil Evans (1957-63)

Davis’ collaboration with the pianist and arranger Gil Evans was legendary. My overall highlights are Miles Ahead, The Duke, Summertime, and Solea.

Miles Ahead (1957)

  • Miles Ahead
  • The Duke
  • Blues for Pablo
  • New Rhumba

Another point: this is not dancing music. I believe that some of Davis’ concerts even had signs up to stop people from dancing. This is in stark contrast to earlier years of jazz, which developed around a culture of dancing halls and highly rhythmic music. Take this with a grain of salt, but it seems like one of the motivations for this era of jazz was to prove to white people that black people were capable of inventing a rich and complex art form that was musically on a par with classical. Indeed, Lee Morgan, the jazz trumpeter, famously said that jazz should be called “black classical music”.

Porgy and Bess (1959)

  • Summertime
  • It Ain’t Necessarily So
  • Buzzard Song

Porgy and Bess is a famous opera by George Gershwin. The Gershwin songbook has been covered many times, and the rendition by Louis Armstrong and Ella Fitzgerald might be the most famous. There have been more arrangements of summertime than anyone can keep track of, but I think Davis’ may well be the best.

Sketches of Spain (1960)

  • Solea
  • Will O’ The Wisp
  • Concierto de Aranjuez: Adiagio
  • The Pan Piper

This album was inspired by his wife, Frances Taylor Davis, and her love of flamenco dancing. Sketches appears to be his first movement toward an acoustic-electric sound. I was surprised to learn that this album is somewhat controversial among Davis fans, so I think it’s thoroughly underrated.

The Second Great Quintet (1964-68)

Miles Davis’ Second Great Quartet had Herbie Hancock on piano, Ron Carter on bass, Tony Williams on drums, and Wayne Shorter on tenor saxophone. Its sound was more unconventional, and arguably the 60s are passed the ‘golden era’ of jazz. E.S.P. and Seven Steps to Heaven are my favourite albums from this era. My overall highlights are Seven Steps to Heaven, E.S.P. and Agitation.   

Seven Steps to Heaven (1963)

  • Seven Steps to Heaven
  • I Fall in Love Too Easily
  • Basin Street Blues

One of my favourite things about jazz is the extent to which its subject matter is universal. Many (most?) popular songs are about love, which is fine, but it’s narrow. The best jazz is a kind of philosophical meditation about the tension between planning and improvisation. The benefits of bottom-up versus top-down design. The extent to which life is an interrelated mesh of trade-offs and constraints. How beauty – and perfection –balances on a knife-edge between order and chaos.

E.S.P (1965)

  • E.S.P.
  • Agitation
  • Mood

I don’t have the musical talent to predict which way a piece of jazz will go, but there is a very satisfying way in which it feels like the notes makes sense after they’re played. It’s curious: I feel like asking ‘how could it have been otherwise?’, when of course, it could easily have been otherwise.

Miles Smiles (1967)

  • Footprints
  • Circle

Bill Evans, who was a pianist with the Miles Davis Quintet for a time, once said “There are no wrong notes, only wrong resolutions.” The way the very same note can sound totally different when played accidentally by an amateur compared to consciously played by a virtuoso is fascinating to me.

Nefertiti (1968)

  • Nefertiti
  • Pinocchio

Nefertiti is one of Davis’ last albums of ‘conventional’ jazz before he developed a more experimental style.

The Fusion Period (1968-91)

One of the key things to understand about Miles Davis is the extent to which he was continuously switching up his style and changing genres. Jazz fusion and Avant Garde jazz are acquired tastes. I found Directions to make for the easiest listening. My favourite songs from this section are Love for Sale, Duran and Black Satin.

Filles De Kilimanjaro (1968)

  • Filles de Kilimanjaro
  • Frelon Brun

The trumpeter Wynton Marsalis once said that “in jazz, every moment is a crisis.” This was a much more elegant way of putting one of my earlier points about the universality of the themes in jazz.

In a Silent Way (1969)

  • Shhh / Peaceful
  • In a Silent Way

Bitches Brew (1970)

  • Bitches Brew
  • Pharaoh’s Dance
  • Miles Runs the Voodoo Down

Perhaps the most famous jazz fusion album of all time. There are single versions and shorter edits of many of these songs, which you might appreciate if you don’t like listening to long songs. Many of his fusion pieces are 30+ minutes long.

A Tribute to Jack Johnson (1971)

  • Yesternow
  • Right Off

On the Corner (1972)

  • On the Corner
  • Black Satin

Miles took a lot of inspiration from world music, as evidenced in Sketches. This album uses a lot of Indian percussion.

Circle in the Round (1979)

  • Circle in the Round
  • Love for Sale
  • Two Bass Hit

Placing this in the fusion period is somewhat of a mischaracterisation, because this album compiles 15 years’ worth of previously unreleased tracks. Nonetheless, the title track is the first recording where Davis used electric instruments.

Directions (1981)

  • Directions I
  • Directions II
  • Duran
  • Water on the Pond

Tutu (1986)

  • Tutu
  • Time after time – live in Nice (from the Deluxe edition)

Davis continued making music until he died in 1991, but the most recent material is of more mixed quality and never really found an audience. Some of the live recordings from this period are much better, however, as we’ll see in the next section.

Live Recordings and Soundtracks

I actually prefer listening to live recordings. They often last much longer than the originals and they give interesting re-imaginings and re-interpretations of familiar tracks. My favourite songs in this section are So What, Walkin’, Générique, Sur l’autoroute and Autumn Leaves.

Ascenseur pour l’échafaud (1958)

  • Générique
  • Sur l’autoroute
  • Florence sur les Champs-Élysées
  • Dîner au motel

This album was an improvised soundtrack for a French film. After spending some time in France in the 50s, Davis was frustrated when he returned to America’s more backward racial attitudes. It’s possible that this anger influenced his music, but I really don’t know.

The Complete Live at the Plugged Nickel (1965)

  • If I Were a Bell
  • Stella by Starlight
  • Yesterdays
  • My Funny Valentine
  • I Thought About You

One of the themes I find most interesting in jazz is the constant tension between improvisation and planning. The different takes sound really different to one another. You would naively think that, however good your music is when you’re composing on the spot, it must be better when you can plan it out in advance. But something is lost when you write the music down. In the very early days of jazz, even the introduction of recording technology was controversial, because when you can record music there is a sense in which it becomes set in stone, and unchanging. But jazz, it was argued, was all about change and revisions. There’s an obvious parallel here with Socrates’ dislike of writing. I say “on the spot” but this is unfair: jazz improvisation requires tremendous practice and intellectual effort. People sometimes conflate “genius” with “very talented”, but so far as I’m concerned, Miles Davis was legitimately a genius in this regard.

“Four” & More (1966)

  • Four
  • So What (this is my favourite take on the song)
  • Joshua
  • Walkin’
  • There is No Greater Love

Stockholm 1960 Complete (1992)

  • Stardust
  • Lover Man
  • Makin’ Whopee
  • Autumn Leaves

Side note: It’s striking how many errors are made in the transcription of lesser-known albums onto Spotify. A lot of the songs on this album have inconsistent capitalisation, i’s that aren’t capitalised, and others have spelling errors.

Miles and Quincy Live at Montreux (1993)

  • Solea
  • Boplicity
  • The Duke
  • Summertime
  • Springsville

This was recorded the year that Davis died, with Quincy Jones. You can tell that he had lost some of his technical proficiency with the trumpet by this point. While Davis was at a time extremely fit and enjoyed boxing, decades of frequent alcohol, cocaine, and heroin use took its toll. With Davis, though, you’re generally not listening for technical proficiency: he certainly couldn’t play faster or higher than some other trumpeters. It’s more that his style is incredibly distinctive. Even today, I’m not aware of any trumpeter that can make their instrument sound the same way Miles could, which is an impressive feat for any musician.  

Bonus: Books, documentaries and films about Miles Davis I recommend

Miles: The Autobiography

Miles Davis wrote an autobiography with Quincy Troupe, which understandably is essential reading for understanding him.

Jazz (Ken Burns)

This is a documentary series and accompanying book. I haven’t been able to find a place to affordably watch the documentary, but I recommend the book wholeheartedly. I also recommend this YouTube interview with Wynton Marsalis and Ken Burns about the series. A quote I liked: “When historians in 1,000 years look back on America, it’ll be remembered for three things: baseball, the constitution, and jazz.”

Miles Ahead

This film is set during Davis’ dormant period in the late 70s in which he was struggling with drug addiction and not making any music. It’s pretty good, though not amazing. The film was Don Cheadle’s directorial debut and stars Don Cheadle and Obi-Wan Kenobi Ewan McGregor. Cheadle did a great job with the raspy voice. People who know about this sort of thing say that the fingering and playing look believable because Cheadle actually learned how to play trumpet for the film.

The film was scored by the excellent Robert Glasper, who also produced the ending track, which is really good. I’ve included the Go Ahead John edit from the film in the playlist. I have had frustrations trying to find jazzy hip hop where the jazz wasn’t just bad or excessively electronic. Kendrick Lamar and Glasper seem like exceptions to this. Glasper also made a video for Wired where he reviewed jazz scenes in films, which you should watch.  

Miles Davis: Birth of the Cool

One of my favourite documentaries. It contains interviews with Herbie Hancock, Wayne Shorter, Frances Taylor Davis, and others. They even interviewed the French woman whom Davis had a relationship while he was living in Paris in the 50s. The soundtrack is really good, and it includes Donna Lee, which I’ve also included in the companion playlist.

Other podcasts and videos

There were interesting discussions about Miles Davis on the Conversations with Tyler episodes with John McWhorter and Kareem Abdul-Jabbar (you can search the transcripts to skip to the relevant part). I’m not sure whether they mention Davis specifically, but Tyler Cowen’s discussions with the music critics Alex Ross and Ted Giola are also excellent. Giola now has a Substack that you should subscribe to if you enjoyed this post. I also enjoyed this video from the YouTube channel Polyphonic about Kind of Blue. All of his other videos about jazz are also worth watching.

Thanks to Sydney, Gytis and Tom for reviewing drafts of this piece.

Book Review: The Signal and the Noise

The Signal and the Noise: The Art and Science of Prediction was written by Nate Silver, a consultant-briefly-turned-poker-player-turned-political-analyst who is most famous for the election forecasting website The Signal and the Noise is one of the small number of books – along with Philip Tetlock’s Superforecasting – that aim to seriously assess the question of how predictable the future is, and how people can systematically improve their prediction ability. This is a question which is of a lot of interest to me. Improving judgements about the future seems to be highly important in many areas (what will the effects of a policy be? When will different technological developments occur? How many people will die from COVID?) and very little attention is paid to it. I found The Signal and the Noise to be thoughtful, and I learned a lot from it.


A running theme of this book is that humans don’t have very good track records predicting the outcomes of complex systems. But one domain where humans have excelled is weather forecasting. Weather forecasts are amazingly accurate relative to the complexity involved. In the mid-70s, the US National Weather Service was off by about 6 degrees (Fahrenheit) when trying to forecast three days in advance. This isn’t much more accurate than what you get if you look at long-term averages – as in, what temperature is most likely in this region at this time of year, not taking into account any specific information. Now, the average miss is 3.5 degrees. This is actually slightly less of an improvement than I would have guessed, although to reduce the error in a forecast by a factor of two requires way more than twice as much effort, since errors can compound.  

I was surprised to learn how large a role humans still play in weather forecasting. Having a human expert use their judgement in assessing many computer-generated forecasts is better than any of the forecasts are by themselves. Humans make precipitation forecasts 25% more accurate than computers alone and temperature forecasts 10% more accurate. Moreover, the accuracy added by humans has not significantly changed over time, so humans have been getting better at the same rate as the machines (!). If you’re wondering why the weather forecasts you use don’t feel very accurate, it’s in part because weather services are private companies that tend to exaggerate forecasts for appeal; you won’t see this inaccuracy in government forecasts. In particular, meteorologists are known to have a “wet bias” – they forecast rain more often than it actually occurs.

There have been some pretty tremendous positive social externalities of commercial weather forecasting, most notably in creating sophisticated early warning systems for extreme weather. The ability to predict typhoons in India and Bangladesh, for instance, has probably saved many thousands of lives. Silver has a few stories in here about people who refuse to leave their homes during an evacuation because of an unjust scepticism of the forecasts. There also appears to be an exposure effect egoing on: studies of hurricanes find that having survived a hurricane before makes you less likely to evacuate future ones. 


The terms ‘fox’ and ‘hedgehog’ used in this book come from the Greek poet Archilochus, who wrote that “a fox knows many things, but a hedgehog knows one big thing”. Foxes are people who don’t have grand unified theories, who constantly revise their beliefs to account for new evidence, and live in uncertainty. Hedgehogs are partisans, and have overarching worldviews which they’ll contort the evidence to fit.

The legendary psychologist Philip Tetlock ran a forecasting tournament in which he tracked and graded the predictions of political experts including professors and government officials over nearly two decades and which he summarised in his book Expert Political Judgement. The main finding: experts are barely more accurate at prediction than chance, and usually perform worse than simple extrapolation algorithms (like “assume nothing will change”). There were too many hedgehogs and not enough foxes. The incentive for pundits and journalists is not to actually be accurate; it’s to appear reasonable while giving novel and entertaining predictions. Indeed, another of Tetlock’s major findings is that the more often an expert was on TV, the less accurate their predictions were.

Tetlock also found an overconfidence effect: when an expert says something has no chance of happening, it happens 15% of the time. When they said it is guaranteed to happen, it happens 75% of the time. While foxes get better at predicting with more information, hedgehogs get worse. If you have grand theories instead of partial explanations, having more facts can make your worldview even less accurate. Partisan differences in prediction were not seen in general (people were relatively unbiased in guessing how many seats republicans vs. democrats would win) but there were marked in specific cases (a left-leaning pundit is much more likely to say a specific democrat will win). These predictions were graded using a Brier Score.

(I wonder if this generalises? If we have some kind of broad philosophical or political worldview that biases us, we might actually see more bias the more we zero in on specific cases. Hence, while talking about specifics and partial explanations is usually the better way to get at the truth, to be effective it might require some deconstructing of one’s prior beliefs.)


The woeful state of prediction might lead you to worry about climate science, where government policy is explicitly shaped by expert forecasts. Indeed, the magnitude of warming from climate change has been overestimated by scientists historically. The actual level of warming was below the 1990 IPCC estimates’ most optimistic projection. In response, the IPCC revised down its models in 1995, and now the observed outcomes fall well within the confidence interval of the projected outcomes (albeit the warming is still slightly less than predicted). You can certainly tell a story here about bias: scientists probably want to find a large warming effect and they think (correctly) that we’re at way more risk of panicking too little than too much. However, these estimates assumed a “business as usual” case; so, one factor that wasn’t addressed adequately was that Chinese industry caused an increase in sulphur dioxide concentration starting around 2000, and sulphur dioxide causes a cooling effect. People forget about the other factors that contribute to warming – I was unaware that water vapour is actually the factor that contributes the most to the greenhouse effect! This all seems complicated to take into consideration so the less-than-stellar prediction performance of climate scientists can probably be forgiven. They also seem to have humility: just 19% of climate scientists think that climate science can do a good job of modelling sea-level rise 50 years from now, for instance. At least as of when this book was published (2012), the effect of climate change on most extreme weather events also appears to be unclear. This is a level of uncertainty that the media definitely fails to communicate.  

Notably, the estimates around climate change are spectacularly noisy, which is well-known, but I think I had failed to appreciate just how noisy they are. Over the last 100 years, temperature has declined in one quarter of decades – for instance, global temperatures fell from 2001 to 2011.

Another thing people seem to forget is for how long we’ve known about the greenhouse effect. It was discovered by Fourier (of Fourier transform fame) in the 1880s, and Arrhenius in 1897 was the first to predict that industrial activity would lead to a warming effect.


The economist John Kenneth Galbraith famously said that “the only function of economic forecasting it to make astrology look respectable.” Indeed, at least in terms of asset pricing, we shouldn’t expect economics to be of any help at all because of the efficient market hypothesis (EMH). This says that stocks and other financial products are priced in such a way that encapsulates the sum total of the information available to the market, such that individual trader advantage is rare. There are two components to EMH, which Richard Thaler sometimes calls the No Free Lunch assumption and the Price is Right assumption. No Free Lunch, or, colourfully, the Groucho Marx theorem, says that you shouldn’t be willing to buy a stock from anyone willing to sell it to you; in other words, it’s difficult if not impossible to consistently beat the market. The Price is Right says that assets are efficiently priced in a way that encapsulates all information.

Thaler has made a career out of exposing the extent to which economic models do not take sufficient account of human irrationality, and he is the ideological arch-nemesis of Eugene Fama, the father of EMH (they’re also golfing partners, which I think is cute). Thaler has a famous paper in which he looks at the company 3Com, which created a separate stock offering for its subsidiary Palm. There was a scheme whereby 3Com stockholders were guaranteed to receive three shares in Palm for every two shares in 3Com that they held, which implied that it was mathematically impossible for Palm stock to trade at more than two thirds of the value of 3Com stock. Yet, for several months, Palm actually traded higher than 3Com, through a combination of hype and transaction costs.

The final point that Silver makes about EMH is that it’s in this fascinating epistemic state where if people actually believed it was true, it would stop being true. The only reason people trade stocks is because they think that they have better judgement than the market (if you invest in a portfolio that tracks the market average you will outperform 50% of traders by definition). This mirrors a lot of what people say about startups: if people actually believed that almost every possible great company idea has already been taken, then they wouldn’t start so many companies, undermining the process that made the original statement close to true.

Why does Silver talk about a theory of asset pricing so much? Because it’s epistemically important to forecasting. If there’s an efficient market for ways to improve the world, then if something were a good idea, someone would already be doing it. If there was an efficient market for ideas, every good idea would already have been tried and rise to the level of scientific consensus. And yet science is subject to massive systemic flaws, and huge opportunities for improving the world remain untapped because of inertia and apathy. Improving our forecasts of the future is important. It seems like a lot of people stand to make a lot of money from doing this. It seems like a small community mostly consisting of nerds on the internet would not be able to massively advance this field. But this impression is wrong.

Silver points out that if you look at the predictions of the Blue Chip Economic Survey and The Survey of Professional Forecasters, the former has some forecasters which do consistently better than others over the long run, but the latter doesn’t. The reason why is that Blue Chip isn’t anonymous, and so forecasters have an incentive to make bold claims that would garner them a lot of esteem if they turned out to be true. One study found a “rational bias” – the lesser the reputation of the institution that someone was forecasting from, the more bold they were in the claims they made. While considerations of esteem probably worsen forecasts overall, they lead some individuals to consistently outperform the crowd.  

All of this should help us to understand bubbles. If EMH is true, how could outside observers notice massive market inefficiencies? Robert Shiller pointed out how the price-earnings ratio (share price divided by earnings per share) during the dot-com boom was unreasonably high, which was the sort of thing that had previously preceded a crash. One of the reasons why the bubble did not sort itself out despite people like Shiller pointing this out is the career incentives of traders: if you bet against the market and the market doesn’t crash, you look like an idiot, while going along with the herd won’t result in exceptionally bad personal outcomes. Silver says there is significant evidence that such herding behaviour exists. 

Given all this volatility, it shocked me to learn that, over the long run, house prices in the US were remarkably stable until recently. In inflation-adjusted terms, $10,000 invested in a home in 1896 would be worth just $10,600 in 1996 (as measured by the Case-Schiller index). The value of such an investment would then almost double between 1996 and 2006!   


There are a lot of interesting applications of the lessons from the science of prediction. One of the most exciting to me is predicting what research is going to replicate. One of the key lessons we should take from The Signal and the Noise is that academics, like everyone else, have all sorts of motivations, including prestige. Through honest motivations, scientists might go along with results that conform to their expectations and worldview, but that a financial market wouldn’t price as being likely to actually be true. While markets have problems (see above), they’re a vast improvement over hearsay and surveys. A ‘prediction market’ works because it actually incentivises people for accuracy in a way they almost never are in other domains. It also works in part because of the wisdom of crowds: group aggregations of forecasts outperform individual ones by 15-20% on average.

Many of you will know this story: John Ionaddis publishes a paper with the provocative title Why Most Published Research Findings Are False which argues that due to the high number of researcher degrees of freedom, and the large variety of results that can be demonstrated with sophisticated statistics, most published research is probably wrong. More than a decade later, he seems to have been proven right. Bayer Labs found that more than two thirds of psychology research papers failed to replicate. Hence, the possible gain from a prediction market in study replication is large. One such project is Replication Markets.


Scott Alexander criticises how people sometimes use the low total death tolls from terrorism as a way to mock conservatives, or people who are concerned about terrorism in general. Most years, lightning kills more people in the US than terrorism, so why worry? Well, here’s a graph of the number of people that atomic bombs have killed since WW2 compared to the number of people who die by lightning each year. Would this be a convincing argument for not worrying about nuclear war? The tail risks are the whole goddamn point.

If you’ve read The Black Swan, you’ll know that lots of things are like this, with ‘heavy-tailed’ risk, and that we sometimes try to shoehorn these into normal distributions.

Earthquakes are distributed according to one such heavy-tailed distribution – a power law – whereby for every one point increase on the Richter scale, an earthquake is ten times less likely. So the bulk of the devastation comes from just a few earthquakes. The Chilean earthquake of 1960, the Alaskan earthquake of 1964, and the Great Sumatra Earthquake of 2004 accounted for half of all energy released by all earthquakes in the world over the entire 20th century! What else is less like height and more like earthquakes?


In one of the book’s middle chapters, Silver uses terminology about infectious disease that many of us have become familiar with over the last couple of months, particularly SIR models. One interesting thing he talked about was the failure of SIR models to account for how there wasn’t a re-emergence of HIV in the early 2000s among active gay communities like that in San Francisco (there was an increase in unprotected sex and other STDs). It’s actually still somewhat a matter of debate why this happened but probably it was because people began to “serosort” – namely, choose partners who had the same HIV status as them. This goes against one of the major assumptions of the SIR model, which is that interactions among individuals are random.

The next few pages blew my mind the most out of anything I had read in a while.  I can’t believe I hadn’t heard of President Ford’s 1976 campaign to vaccinate 200 million people against a suspected H1N1 pandemic. The vaccine dramatically increased the rates of the nerve-damaging Guillain-Barré syndrome, and the public turned against it, such that only 50% of people were willing to be vaccinated! The severity of the outbreak also turned out to have been exaggerated, so the government gave up after 25% of people were immunised. How have I not seen this being brought up in the context of COVID?


I recommend this book, particularly if you’re not already familiar with Philip Tetlock, forecasting, and Bayesian statistics. For people who are already interested in that kind of thing, I can still recommend skimming. I’m sure I’ll write about forecasting again on this blog at some point – I didn’t even have time to talk about superforecasters, the people who can consistently outperform expert predictions. ★★★★☆

Afterword: Philosophical Pondering on the Problem of Prediction

One question that Tetlock sometimes gets asked about is whether it’s nonsensical to ascribe a probability to an event that only occurs once. If you think the universe is deterministic, you might say that the probability of a certain candidate winning an election is either 0% or 100%, but you simply do not know which. So, in what sense can this be evaluated probabilistically? Does probability represent something metaphysical about what the outcomes would be if the trials were run infinitely many times? Or just a degree of belief? The former view is identified with the frequentist school and the latter the Bayesian school. Regardless of one’s philosophical position, Tetlock’s approach is to just get on with it: if we look at the set of all supposedly unpredictable things, do the things you predicted would happen 10% of the time happen 10% of the time?

Viewing probability as just degree of belief is actually very counterintuitive. There are problems with this that I still haven’t figured out, like the distinction between external and internal ‘credences’, of degrees of belief. I may think there is a 50% chance that Trump will win re-election, but isn’t there some higher-order uncertainty I have about whether I’m using the correct mental model to assess this, or whether I’m actually in a computer simulation, or something? But doesn’t this eliminate the initial theoretical appeal of having all considerations cash out into a single credence? What if your credences hold some mathematically impossible property?

David Hume thought that, because we don’t have certainty, saying that the sun will rise tomorrow is inherently not any more rational than saying it won’t. More recently, probability-as-beliefs was famously opposed by the statistician Ronald Fisher. One of the main problems with the frequentist school is how much of forecasting and probability turns into a game of finding the correct reference class, or relevant comparison group. The reference class for the die you roll is fair six-sided dice, but what reference class would the 9/11 attacks be in, for instance? So the principal objection to the Bayesian approach – that it is too subjective – applies also to views of probability-as-frequency.

Do All Languages Communicate at the Same Rate?

The speakers of some languages have a reputation for talking quickly while others have a reputation for talking slowly. Is this because some cultures actually communicate quicker than others, or are they just using more words to communicate the same information?

That is what Coupé et al. aims to answer empirically, using a dataset of 17 languages. They conclude that, indeed, the information communicated per second of speech is similar across languages – in particular, around 39 bits/s.

Duration of time that speakers of various languages require to communicate the same information

If you’re wondering how a syllable could have a bit value, you have a fun afternoon ahead of you reading about information theory. In an information-dense 8 bit/syllable language like Vietnamese, you on average have a 1 in 256 chance of predicting the next syllable, while in a 5 bit/syllable language like Japanese, you have a 1 in 32 chance. So, there are two competing strategies that balance each other out: either have information-dense syllables, and speak more slowly, or have information-sparse syllables, and speak quicker. Cool!

I couldn’t find anything on whether the same thing is true of dialects. Speakers with some accents speak much more quickly than others – for instance, Northerners are known for speaking quicker than Southerners in the US. This would be a great secondary school or undergraduate summer project. And it wouldn’t even be that hard to run – you could just give people the same piece of text to “translate” into their dialect, then see if the same information is communicated in the same amount of time.

Can a language benefit its speakers in general?

Researching speech efficiency got me wondering: are there specific benefits that speaking a certain language can give you?

I’m somewhat biased against this being true. If speaking East Whereverese raised your IQ by 5 points, you’d expect its beneficial features to be adopted by other languages, or be independently converged upon by multiple languages.

In Outliers, Malcolm Gladwell suggests that one of the reasons why Chinese students perform so well on mathematics tests is because of beneficial features of the Chinese language itself: namely, saying words in English is clunky and takes a lot of syllables, but numbers in Chinese are quicker and easier to speak. Being able to hold numbers in working memory for even a fraction of a second longer can be a real boon for maths ability. By age 4, Chinese children can already count to 40, while Americans of the same age can only count to 15, putting them a full year behind. And this is before formal education starts, so you can’t pin this one on the education system.

Again, this is mysterious to me. If you can upgrade your children’s numerical reasoning by a full year just by giving them a vocabulary that makes it easier to talk about words, why haven’t other languages evolved to do the same thing? I’ve always thought those parents that teach fake languages to their kids were wasting their time, but if Klingon is anywhere near as efficient in talking about numbers as Chinese, maybe they have a point! Are there other cognitive capabilities that the English language enhances that compensate for the mathematical disadvantage it puts its speakers at?

(Note: Scott Alexander points out how Chinese mathematics test scores are kind of fake, in that China struck a deal whereby it only administer the PISA exam (the standard test for comparing countries’ mathematics ability) in its most educated provinces (!). If the US were allowed to do the same thing, you find that the Chinese mathematical advantage goes away. But this wouldn’t explain Chinese kids being able to count higher. So, all in all, not sure what to make of this.)

Appendix: Podcast listening speed

It still boggles my mind that not everyone has realised that you can listen to podcasts and audiobooks at >160% speed while losing no comprehension. This seems to imply that the rate-limiting part of speech – and therefore why languages communicate at the same rate – is how fast people can speak, not how fast people can listen. In which case, I await the day where I can have a neural implant that allows people to communicate all the same information to me in half the time 🙂

Follow-up on the Open Borders Review

Appendix A: What I Left Out

There’s a lot more to talk about with this book, but my main review has all the points I have a strong view about. Caplan discusses the objection that immigrants would lower average IQ, and talks about Garrett Jones’ book Hive Mind, which argues that national IQ is very important in determining national prosperity. I’ll point you to the Slate Star Codex review for more discussion (with the caveat that IQ research is very controversial so please don’t get mad at me).

Caplan has a cute section of the book called ‘All Roads Lead to Open Borders’ in which he describes how open borders remain a good idea under a wide variety of philosophies, including utilitarianism, Kantianism, Christianity, and libertarianism. I don’t have strong views on the philosophy of immigration – my impression studying the subject at university was that any a priori objections should only make a difference at the margin, and are dwarfed by the empirics. Still, there are some plausible moral values you could hold that would make immigration less appealing – such as cultural preservation, or countries’ right to self-determination. Conversely, if you especially value cultural diversity or individuals’ rights to freedom of association, open borders seem more appealing.

One thing I didn’t talk about was brain drain, mostly because I don’t think it’s a very significant problem. Developing countries are not even close to coming up against the ceiling of people who are capable of doing in-demand jobs like being health professionals. Crudely, the reason why there aren’t many engineers in Chad isn’t that Chad trained a bunch of engineers who all left; it’s that Chad doesn’t have many engineers full stop. A lot of the philosophical literature on open borders also seems to be confused about this point. Doctors immigrating from developing countries doesn’t reduce the supply of doctors in those countries. The Philippines’ supply of nurses has actually increased as a result of the fact that they send so many nurses abroad. 

Appendix B: Arguments that Caplan Didn’t Use

  • How much of India and China’s economic growth is as a result of the fact that they’re really big, and therefore, moving across them is a lot like immigrating? When Caplan pointed this out, I was pretty surprised I hadn’t thought about it before. Were the two major economies to take drastic steps in reducing poverty in recent decades able to do so largely because they’re really big? This is like one panel on one page, but I felt like he could have developed the argument more. In general, I think the book’s argumentative style leans too highly on Estimates by Economists™ and not enough on case studies and natural experiments. Do more populous countries have greater growth in the long run? If so, this points us in the direction of open borders. Relatedly, I liked how Caplan talked about what Lant Pritchett calls ‘zombie economies’ – economies kept alive by restrictions that forbid people from leaving. A shockingly large share of the US has been declining in population for decades, yet we would regard it as absurd to say that people shouldn’t be allowed to leave Nebraska because doing so would go against Nebraska’s interests.
  • There are various arguments related to long-termism that Caplan didn’t use; namely, the downsides of immigration (higher crime, perhaps draining the government’s budget) are temporary but the upsides (higher economic growth) bear their fruit over centuries and will likely affect billions of future people. If you buy the argument, popular within effective altruism, that what matters most morally is our consequences on the long-run future, this would seem to be a point for open borders.
  • Caplan makes it seem like it’s an open-and-shut case that immigration doesn’t lead to an increase in unemployment. But many economists are also fans of the minimum wage. But surely there’s a tension here? If the minimum wage has no disemployment effects, the labour market is perfectly inelastic, and if immigration has no disemployment effects, the labour market is perfectly elastic. So how elastic is the labour market?
  • Such a disproportionate amount of innovation comes from immigrants. More inventors immigrated to the US from 2000 to 2010 than to all other countries combined. Immigrants account for a quarter of total US invention and entrepreneurship. Maybe this is just because America disproportionately lets smart and innovative people move there. But maybe there are some agglomeration effects going on here specifically related to immigration? Immigration – or more particularly, clustering people together – seems to have been key to the success of various intellectual hubs throughout history, like the Bay Area recently, Vienna in the 20th century, and Edinburgh in the 18th century. This seems like a ripe topic for progress studies to tackle. Aesthetically, I agree with Caplan’s choice not to talk about this much. People talking about all the “amazing contributions made by [insert immigrant group]” often comes off as condescending, in much the same way as token engagement with other cultures might. Make the case for immigration from prosperity and freedom, or don’t make it at all! But it still has to be confessed that immigrants do seem to contribute a disproportionate amount – technologically, artistically, scientifically, and culturally – to the US.
  • I think there are good reasons to believe that way fewer immigrants would actually move than Caplan presupposes. During the entire Greek financial crisis, only 3% of the Greek population moved country (!), at a time when the unemployment rate was 27% – and remember, Greeks have more than a dozen prosperous destination countries to choose from with no paperwork involved! Inertia is the most powerful force in the universe. Caplan’s defence of his high implicit estimates is that, once the ball gets rolling, more and more immigrants from a particular country will move – for instance, historically immigration from Puerto Rico to the US was lower than you would expect given the difference in economic opportunity, but then Puerto Rican communities formed in many US cities, and more and more people moved. Gallup finds more than 100 million people want to migrate to the US. 750 million say that they would leave their home country if they could. But we have reason to doubt people would actually act on this. This makes open borders a little more palatable to people that are sceptical of immigration: it wouldn’t be as different to the status quo as you might expect. 
  • What factors have led Canada and Australia to handle immigration so well?
  • There is a general perception that Muslim immigration to the EU has gone poorly. How much of this is hysteria? Why would future rounds of immigration not have problems in the same way?
  • Greying is not something that Caplan talked about much. At first this might seem surprising – one common-sense case for immigration is that people in Europe and America are getting too old to work and they need immigrants to replenish their workforce.
  • There aren’t really jobs that “Americans won’t do”, since, if people don’t like doing something, the wages will rise until they start doing it to meet demand. However, this price is such that there’s significant deadweight loss – mutually beneficial trades that no longer occur. For instance, more people would get more childcare if the government allowed more immigration. Caplan discusses this, but I didn’t feel sufficiently inspired to think about how this would be great for me personally. If I lived in a place with open borders, I’d probably hire a personal assistant or something.

Appendix C: One Billion Americans

Another book I read recently and recommend is Matt Ygelsias’ One Billion Americans: The Case for Thinking Bigger (Caplan reviewed it on his blog). He argues for large-scale population growth, partially through immigration but mostly through an increase in fertility, to maintain American pre-eminence over China and India. He argues that, for all of its failings, American dominance is better than the alternative. And America is at a disadvantage on this front by having a billion fewer people than the Asian giants.

I’m not sure this argument should have gone in the book – it would take a long time to justify, and open borders appeals to a lot of left-libertarian sensibilities that might be offended at the idea of American global hegemony. But it would be an interesting project for the open borders community to look at the geopolitics of population growth. How important are marginal increases in population to geopolitical power? Are spurts in population growth followed by increases in various measures of hard power? Soft power?

Book Review: Open Borders

Bryan Caplan is an economist at George Mason University and all-around interesting guy who is known for his out-there views about various social and political issues (especially education). Open Borders: The Science and Ethics of Immigration is his latest book, which argues for an end to all restrictions on migration and is in the format of a graphic novel illustrated by Zach Weinersmith of SMBC fame. The first thing I would say about this book is that the graphic novel format works really well. The art style is cute and I think graphic novels are heavily underrated. Realistically, most people are not going to read a regular book about the economics of immigration. But this way Caplan can lure us in with fun cartoons! The next thing I would say is that the book makes an important argument on an issue where people have particularly poorly thought-out opinions. The data are pretty clear that immigration is massively more beneficial than most people realise – certainly economically, and perhaps socially too. However, upon reflection there are serious objections to open borders, and the arguments in the book have a number of omissions.

The argument

Caplan really does believe that there should be no restrictions on immigration whatsoever, and that’s exactly what his cartoon representation in this book argues for. The basic argument goes like this: people should, in general, be allowed to make decisions that they think will improve their lives, assuming they’re not hurting anyone. Moving to a new country is exactly such a decision. Since immigrants often move in search of work, moving is associated with a massive increase in economic prosperity: by moving to the US and receiving no additional training or education, the average citizen of a developing country can expect their income to increase fivefold; for countries like Nigeria, the figure is tenfold. This is because developed countries are safer, more prosperous and have better quality institutions, so immigrants are more productive in them. The gains are so vast that a standard estimate is that open borders would double world GDP. And yet rich countries continue to restrict immigration, sometimes through formal caps, and sometimes through complicated bureaucracy and paperwork which at best dissuades people from entering and at worst makes it literally impossible (like rejecting you for not filling out the middle name section on a form when you don’t have a middle name). Some of the arguments against immigration are xenophobic or racist, but many are legitimate concerns bought up in good faith. However, most (all?) of these are simply not borne out by careful consideration of the evidence. The consensus among economists is that immigration does not generally decrease natives’ wages. Nor does it lead to an increase in poverty, crime, or a significant strain on the welfare state and social services. While the data about this is more unclear, immigrants seem to be barely different from natives in their political views and they adopt a lot of the cultural values of their destination country. Hence, the contrary considerations are not enough to overwhelm our initial presumption in favour of allowing people to move and massively improve their standard of living, and so we should have open borders.

The objections  

1: Parochialism

Open Borders is an extremely US-centric book. As someone from the land of ‘not America’, this is something that frustrates me about a lot of non-fiction. Caplan justifies his focus on the US by saying that his audience is mostly Americans and that that’s where the highest quality data exists for. But in this case, the book makes a way narrower argument than is set out in the book’s intro. By focusing primarily on America, the case is made stronger than it otherwise would be. For instance, immigrants commit more crimes than native-born Europeans but fewer crimes than native-born Americans. Immigrants to the US also seem to assimilate unusually well (although some people say this is just because European countries are more regulated, and in their infinite wisdom make decisions like forbidding refugees from getting jobs).

Focusing so much on the US is bizarre because the European Union has open borders between its member states! Surely analysing whether this has gone well should be the most convincing piece of evidence about open borders. Ireland is 17% foreign-born, a significantly higher proportion than the US, and from eyeballing the data is looks like the immigration rate to Ireland has nearly quadrupled in the last 20 years. This would seem like a major success story of immigration. Meanwhile, Caplan only talks about the EU for a few panels toward the end of the book. The considerations above don’t seem to justify anywhere near this level of parochialism.

Until the 1920s, the US had de facto open borders, and this is another thing that I wish Caplan had dug into more. It certainly seems like America benefited a lot from immigration at this time (or, at the very least, that immigrant groups like the Irish did) but have people studied what the effects of open borders actually were?

2: Humility

Open borders would be the largest social transformation possibly ever, and there isn’t even that much research about it. We should in general be extremely humble about the prospects that our views about complex topics are completely right, and the downsides from open borders, if we are wrong, could be quite significant.

Caplan is unusually scrupulous at making sure his claims are backed up by the data. His book The Case Against Education is one of the most meticulously researched books I have ever read. So, it was a bit disappointing that there weren’t more margins of error attached to his claims. How confident are we that open borders would really double world GDP? 10%? 50%? 90%? Even with such error bars, after reading about the replication problems in economics and the colourful uses of statistics to get one’s desired conclusion, I don’t find these kinds of projections very convincing compared to natural experiments and case studies, and I mentioned that the EU, the most compelling such example, is not talked about much.

3: Environment

Unless I’m mistaken, at no point does Caplan address the environmental harms of open borders. Moving people from low-emitting poor countries to high-emitting rich countries would lead to a pretty dramatic acceleration in global CO2 emissions. Admittedly, “keep most of the world poor” is a terrible climate change strategy, but there are some climate problems you might want to solve first before advocating for open borders. A world with open borders would be much richer, and so would have a lot more money to throw at the problem of climate change, but how much more would it throw? If the case for open borders were airtight, it would have to address this. I’m confident that Caplan has reflected and come to the conclusion that there are no climate problems that we can solve in a short-enough period of time to justify the harm caused by delaying open borders, but he doesn’t show his work.

Sometimes, climate change gets used as an excuse for opposing almost any societal progress. This is unfortunate. But “open borders would create this gigantic problem, namely massively accelerated climate change, but the benefits outweigh the harms” was not the argument I got from the book. “Open borders are so good, and the objections are not that significant” was the argument I took away from the book.

There are considerations I can think of that would make the environmental objection less serious. Immigration would probably accelerate the trend of urbanisation, and cities are better for the environment (smaller houses, more use of public transportation, etc.). People would also be able to move away from the regions that are worst affected.

I’m also seriously concerned about the animal suffering that would be induced by open borders. I think that we should give a high degree of moral consideration to complex animals like cows and pigs, and that globally, eating meat, 90% of which comes from factory farms, creates an almost unimaginable level of suffering. There are a couple of reasons why open borders would make this worse: the Western diet is more meat-heavy than diets from other places, and richer people in general consume more animal protein. Some people talk about the meat-eater problem: many interventions in global development look much less cost-effective if you give moral concern to animals (since, if the interventions save human lives or make people better off, they lead to greater meat consumption). The high demand may further entrench factory farming as the default way meat is produced. This is not a consideration that most people have when thinking about open borders, but the premises are relatively uncontroversial. Virtually everyone agrees that animals are worthy of moral concern, and many (most?) people see some problem with eating factory-farmed meat, even if they do not act on their discomfort.

4: Culture

Caplan has a section where he addresses the political effects of immigrants, largely drawing on data from Alex Nowrasteh at the Cato Institute finding that immigrants are a tiny bit more liberal than the general population but that their kids and grandkids regress to the political mainstream. Immigrants and natives didn’t have a partisan difference until the 1980s, and the partisan difference comes from immigrants being more likely to identify as independent, not from being more likely to identify as Democrat (although maybe after a while immigrants become acclimated and realise that third-parties never win…). This is interesting but doesn’t address the tail risk of immigration leading to a Trump/Brexit dysfunctional level of polarisation or backlash (admittedly, that would be very speculative). It may be the case that the biggest harms from immigration come from people irrationally freaking out about immigration, but, uh, people are in fact irrational.

Here’s Michael Huemer, in one of the most well-known philosophical defences of open borders, on the effects of immigration on culture:

“Empirically, it is doubtful whether apprehensions about the demise of American culture are warranted. Around the world, American culture, and Western culture more generally, have shown a robustness that prompts more concern about the ability of other cultures to survive influence from the West than vice versa. For example, Coca-Cola now sells its products in over 200 countries around the world, with the average human being on Earth drinking 4.8 gallons of Coke per year. McDonald’s operates more than 32,000 restaurants in over 100 countries.”

This seems to kind of sidestep the objection. Mass migration to the US is not a concern because Coca-Cola will go out of business; it’s a concern because democracy, freedom of speech, and the rights of women and homosexuals are deeply unpopular in much of the world. Importing millions of people from autocracies and societies that are otherwise deeply illiberal may well have adverse effects on democracy. This makes the case for having long waiting times for citizenship pretty good.

The selection effects right now for immigration to the US are really strong, but we have every reason to believe that they would decline under open borders. If immigration restrictions were lifted, the average quality of immigrant would almost certainly drop. This is something Caplan admits to, but the response to it didn’t feel convincing. Just how much selection bias is there in who gets admitted and who doesn’t?

5: Inequality

Caplan is an economist, so I can’t really argue with his reasoning about the economics of immigration. While the book is pretty convincing in arguing that immigration is the best tool we have for reducing poverty in an absolute sense, I’m less clear about the effects on poverty in a relative sense. Poor Americans still have it great by global standards, but they certainly don’t feel that way, and the point of all this prosperity is presumably to make people subjectively better off. Defeating bona fide poverty – the type where people can’t feed their kids – is priority number one, but still!

Currently, the people who move from poor countries to rich countries are self-selected for being hard-working, intelligent, and conscientiousness. But what happens when the really unmotivated ne’er-do-well’s start coming too? Under the current regime, these people would be relegated to the fringes of society. Could open borders even make some immigrants worse off, even if their pay cheque triples?

Caplan also doesn’t really consider the extent to which racism and xenophobia might flare up in response to immigration (though he does have a great section covering the effects on social trust). The countries that are the closest to having open borders are the gulf states; they have many migrant workers from countries like Bangladesh and Sri Lanka. On one level, this is great: Qatar benefits from cheap infrastructure, the Sri Lankans benefits by getting higher-paid jobs. But I do also fear that this will lead to a kind of racially segregated dystopia.

In fact, immigrant groups would be largely stratified based on how wealthy they were to begin with. African immigrants would likely be deeply poor, followed by not-as-poor Indians, then richer Chinese, and so on. What happens to the politics and culture of a society that is that racially stratified? This is of course also a problem now, but I wonder what it might mean to scale it up so much. The fact that levels of education and training correlate with immigrants’ ethnicity vis-à-vis the differences in wealth among countries would lead to a problematic level of statistical discrimination, at the very least.

I initially was very sympathetic to the view – defended by some philosophers – that wealth inequality is not a problem per se; poverty is. But the more I think about it, the more this feels like squabbles over semantics. Yes, the distribution of resources is not intrinsically morally significant, but the mere fact that poor people don’t have very much money isn’t morally significant either. Conducting research about this is hard, and take the literature with a grain of salt, but, holding poverty constant, inequality seems to have lots of negative effects on all sorts of outcomes, including crime. So, given that it has negative outcomes, and is frequently caused by unjust social conditions, inequality – which would be increased within countries by open borders – is worth worrying about!

(Finally, regarding the welfare state, because I didn’t know what section to put it in. One of the more sophisticated considerations contra redistribution is that excessive transfer payments aren’t really compatible with high levels of immigration (unless you want to go bankrupt), and we know with a high degree of certainty that immigration is better at reducing poverty than government programs. But does this actually happen? Do places that grow their welfare state subsequently shrink their level of immigration, or shift it toward higher-skilled immigrants? Is there something funky going on such that support for immigration and welfare became tightly correlated beliefs?)  


Toward the end of the book Caplan discusses whether it’s a good idea to be advocating for open borders, or whether the idea is so radical that it will turn people off immigration even more. He comes to the conclusion that discussing open borders shifts the Overton window toward increasing immigration. I’m not so sure. For how important it is to convince people about things, I’ve seen remarkably little empirical research as to how you do it. Putting group polarisation aside, is it a good idea to give someone a stronger case or a weaker case to convince them to move their views in the direction of the argument?

This book made me think about what low-hanging fruit might exist in the space of increasing immigration. As I mentioned, immigration in many countries is not formally capped but is de facto limited by being confusing and costly. Have people tried to start companies to fill this niche of streamlining immigration? Are there any foundations willing to run this kind of thing as a non-profit? Google turns up surprisingly few results.   

All in all, I recommend this book. The thing it changed my mind the most about is the extent to which wealth is a function of where you are, not who you are. One estimate is that 60-70% of the global wealth disparity is explained by location alone. You could fix the institutions of poor countries from the ground up – but we don’t know how to do this, it would take a long time, and it’s unclear to what extent their problems arise from geography, so wouldn’t get fixed by better policies anyway. Hence, the case for more immigration still looks pretty watertight. I hope to see these arguments developed further!

Preferences for Masculine or Feminine Faces May Be WEIRD

One of the critiques commonly levelled against psychology is that its samples mostly come from Western, Educated, Industrialised, Rich and Democratic (WEIRD) societies. Joseph Heinrich and others published a highly cited paper in 2010 in which they found that the cross-cultural range of psychological variation was much larger than previously assumed, and that WEIRD samples are actually some of the least representative of humans in general. You cannot test a bunch of Yale sophomores and make conclusions about the universal human condition, as it turns out. A bunch of cross-cultural psychology work ensued. Here’s a highlight from Scott et al. 2014:

“A large literature proposes that preferences for exaggerated sex typicality in human faces (masculinity/femininity) reflect a long evolutionary history of sexual and social selection. This proposal implies that dimorphism was important to judgments of attractiveness and personality in ancestral environments. It is difficult to evaluate, however, because most available data come from large-scale, industrialized, urban populations. Here, we report the results for 12 populations with very diverse levels of economic development. Surprisingly, preferences for exaggerated sex-specific traits are only found in the novel, highly developed environments. Similarly, perceptions that masculine males look aggressive increase strongly with development and, specifically, urbanization. These data challenge the hypothesis that facial dimorphism was an important ancestral signal of heritable mate value. One possibility is that highly developed environments provide novel opportunities to discern relationships between facial traits and behavior by exposing individuals to large numbers of unfamiliar faces, revealing patterns too subtle to detect with smaller samples.”

I take this to mean that in the ancestral environment, it may well have been the case that more masculine men made better mates, but the sample size of men that the average woman had observations of was so small that she couldn’t make the inference. And vice versa for femininity. A number of books have given me the general picture that humans’ preferences and dispositions (for instance, boys’ preferences for rough and tumble play) are mostly genetic and don’t result from social conditioning. Explanations like this, that appeal to the group dynamics of interacting with many more people than we did as hunter-gatherers, seem more convincing to me.