Progressive Myths Book Review
A book review and an audit
1 Overview
Everything
has ever published has been good (I’ve read it all). The writing is clear and hilarious, the arguments never have obvious errors, and they’re often convincing. I was therefore excited to read his new book Progressive Myths, which he very generously sent me a free copy of. I was curious to see how much of it I agreed with, as a left of center person who finds Huemer often correct.I can gladly report that the book is very good. Huemer criticizes lots and lots of progressive sacred cows—the sorts of things that you hear floating around in progressive spaces but never really have the time or desire to investigate. Huemer argues conclusively that many of these are false. For instance, there is no epidemic of police killing unarmed black men, seeing as only 32 unarmed black men were killed by police in 2019, and this is pretty typical across years (of course, this is perfectly consistent with there being other serious problems in the criminal justice system). Something that kills 32 people is not an epidemic, seeing as it affects less than one person per state. One’s way more likely to be killed by alligators or lightning.
Most of the things that Huemer says are convincing and are things I’ve heard from other smart people who have investigated them. For instance, it’s often claimed that the reason minorities and women lag behind in various fields is that they’re worried about confirming negative stereotypes about their groups. This phenomenon is known as stereotype threat. Huemer cites a wealth of research showing that this phenomenon is either nonexistent or very small. The fact that there are a bunch of studies finding that it real isn’t very convincing when there are strong academic incentives to churn out P-hacked studies finding—surprise!—left-wing dogmas are true.
One thing that I appreciated about the book is that Huemer admitted when progressives were right. While many conservatives claim that Chauvin didn’t really kill Floyd or that Breonna Taylor wasn’t really murdered or that Eric Garner wasn’t really killed by Daniel Pantaleo, Huemer is willing to, in a book dedicated to bashing progressive myths, admit when the progressives are right. Similarly, in a chapter that debunks the myth that a quarter of women are raped on college campuses, he admits that this might be reasonably close to the number of women who are sexually assaulted on college campuses. I’d looked into this before, and it seems either correct or reasonably close to being correct.
One particularly perceptive point that Huemer made when talking about transgenderism, was that whether particular trans women count as women might be genuinely underdetermined. Given that trans advocates and their critics agree in the case of whether 99% of people are women, it may be that our linguistic categories just aren’t finely grained enough to determine whether trans women are women. So the answer to the question might be vague or indeterminate.
A non-exhaustive list of the myths Huemer seeks to debunk, followed, after the colon, by a very condensed summary of them:
Trayvon Martin was unjustly murdered: no, he attacked Zimmerman and reached for his gun, before being shot).
Michael Brown was unjustly murdered: no, he was reading for Wilson’s gun.
Jacob Blake was unjustly shot: no, he was getting into the car with a knife to try to kidnap his children after the cops told him to stop repeatedly and tased him, and he was moving to stab the officer.
Kyle Rittenhouse is a murderer: no, all the people he shot were brutally attacking him and/or trying to shoot him.
Police shootings are racially biased: no, per stop they are less likely to shoot a black person, and this is also reflected in lab experiments.
Stereotype threat is real and significant: no, the studies supporting it don’t replicate.
The drug war had racist intentions: it was supported by almost everyone including being proposed by many black leaders.
Implicit bias is a reliable measure of racism: no, it’s a better measure of a perception that a group is discriminated against. If you think that black people are mistreated, it’s easier to associate them with bad things.
Women very rarely lie about being sexually assaulted: we don’t have good data on this and the claims that we have data showing women almost never lie are bullshit.
Regulations aren’t extremely harmful to growth: they actually reduce it by about 2% per year.
Masks aren’t clearly unambiguously good for slowing down the spread of disease: their net effect is unclear, probably slightly negative for cloth masks and slightly positive for other kinds of masks.
The book has too many claims for it to be fruitful to fact-check all of them. Most of the ones that I’d investigated—e.g. the claims about Rittenhouse and Blake—were true. So I decided that for this review, I’d pick 2 claims to investigate: the claims about the Trayvon Martin shooting and regulation.
2 Regulations
The most shocking claim I came across in Huemer’s book was the claim that regulation added to the U.S. since 1949 have reduced economic growth by about 2% per year. This is the claim of a 2013 study by Dawson and Seater, and Huemer says it’s in line with other estimates. If this is true, then had there been no regulations between 1949 and 2011, the GDP at the end of 2011 would have been almost 3 times as high as it actually was. If this is right then the average family income in 2011 would have been about 150,000 dollars, rather than about 50,000 dollars. Huge if true!
This means that we would all be about 3 times as rich if it weren’t for high levels of regulation!
Huemer admits this might be an overestimate as “Perhaps, as GDP grew to much higher levels, growth would have plateaued; in other words, perhaps the “2 percentage point” estimate ceases to be valid for societies with extremely high productivity.” He also claims that regulations are primarily supported by big businesses to enable them to strangle small business. Still, even if it’s a bit of an overestimate, it’s still wildly hugely significant.
I tracked down the 2013 study by Dawson and Seater through incredible investigative acrobatics for which I should get a Pulitzer (I googled 2013 study Dawson and Seater and it was the first thing to pop up). It was published in the highly ranked Journal of Economic Growth, so doesn’t seem obviously bogus. The study uses some decently credible indexes of economic growth and then runs a regression which is a way of analyzing causation between two variables. So the method looks pretty legit. They then compare their results to a bunch of results from other studies, and their results seem to be in line with the other studies.
I’m not equipped to seriously analyze statistics. I don’t know if there’s some obvious huge problem with the data. People like Tabarrok and Goldschlag criticize it on the grounds that “time-series evidence from one country could be subject to considerable biases and can be interpreted as causal only with strong assumption.” If I’m understanding this correctly, what this means is that just looking at if growth goes down after regulations are adopted and controlling for a few things doesn’t tell you conclusively if the regulations really are slowing down growth. Seems plausible, but the Dawson and Seater paper is still decently good evidence (regression analyses tend to be regarded as decent evidence). The fact that this is a reasonable middle ground among the various studies means the conclusion is decently plausible—at the very least not obviously wrong, though I’d want to hear from more statistically informed people if there are subtle statistical errors.
This is pretty alarming. If it’s right, it provides an extremely strong case for supporting conservatives on economic policy. If regulations are responsible for more than cutting the GDP in half, then getting rid of them is massively important.
3 Trayvon Martin
3.1 Huemer’s telling of events
Here is Huemer’s summary of the Trayvon Martin shooting.
Trayvon Martin, a black 17-year-old was walking through a neighborhood when spotted by George Zimmerman. Zimmerman thought that Martin was suspicious and began following him, after calling the police. When the police arrived, they found Martin dead. Zimmerman claimed Martin had attacked him and that he’d acted in self-defense.
Was Zimmerman a racist? NBC News thought so, having played a clip of him saying “This guy looks like he’s up to no good. He looks black.” But in fact, this was two statements spliced together uttered at very different times—the second was in response to the police dispatcher asking for Martin’s race. Additionally, while race may have played a role in Zimmerman following Martin, this was likely because in the previous weeks there had been multiple break-ins by black men in the neighborhood. So it’s no more objectionable racial profiling than it would be if after knowing that a 6’5 man committed several crimes, you called the police after seeing a 6’5 man who didn’t live in the area walking around, looking at the houses, at one point fleeing to be between two of the houses (all of which Martin was doing).
The police didn’t initially charge Zimmerman because they didn’t have an adequate legal basis to do so. This wasn’t because of racism but because of the normal police practice of not arresting people until they have an adequate basis to arrest them.
Zimmerman’s telling of events is that after he hung up with police, Martin went up to him asked “Yo, you got a problem?” and when Zimmerman replied “No, I don’t have a problem, man.” Martin replied “well, you got a problem now,” and attacked Zimmerman. Martin got on top of Zimmerman and began slamming his head into the concrete sidewalk and punching him. When Zimmerman called for help, a neighbor appeared, said he’d call 911, but then retreated. Martin, after seeing Zimmerman’s gun, said “You’re gonna die tonight, motherfucker,” and reached for Zimmerman’s gun, at which point Zimmerman shot him.
Huemer argues that all the evidence supports Zimmerman’s story. In the 911 call—which we have a recording of—someone’s screams can be heard in the background which Zimmerman’s family says are from him. Martin’s father initially said the screams weren’t from his son, before later recanting and agreeing with Martin’s mother. Additionally, a third party witness said he saw Martin on top of Zimmerman, beating him up, while Zimmerman called for help.
This is also corroborated by the forensic evidence. A forensic expert in the trial testified that the gun was pressed up against Martin’s shirt but was 2-4 inches away from his skin when it was shot, which makes sense if it was hanging off him, but makes no sense if Zimmerman was on top.
Huemer admits that Zimmerman acted recklessly. Following Martin was a very stupid thing to do. But doing something stupid doesn’t make you a murderer.
The main witness for the prosecution was Rachel Jeantel and her story is the most bizarre. A letter signed with the name Diamond Eugene was presented during the trial, in which the author of the letter said she was on the phone with Martin when he died. She said she heard Trayvon say “why are you following me?” before the phone cut out. During the trial, Rachel Jeantel testified that she had written the note and was Martin’s girlfriend, having heard Martin say “get off, get off.” If this had happened, it would mean Zimmerman was on top of Martin, proving his testimony to be a lie.
However, Huemer argues it’s very likely that Jeantel was an extreme liar and didn’t actually write the note. The evidence for this:
The note was written in cursive. But Jeantel couldn’t read cursive. She then claimed that she had a friend write the letter for her but signed her name (the signature being printed). This is obviously very implausible.
The note was, as previously noted, signed Diamond Eugene, rather than Rachel Jeantel. Jeantel said that was her nickname, which would be both a very strange and surprising nickname, and it’s unlikely she’d sign her name using her nickname.
Joel Gilbert—who is a crank conspiracy theorist, but might be right here—found Martin’s phone records and identified a real woman named Brittany Diamond Eugene who Martin had been talking to when he died. He provides a recording of Diamond Eugene’s voice made by Martin’s attorney and played for the media which sounds nothing like Jeantel and exactly like Brittany Diamond Eugene. Additionally, the attorney had stated that Diamond Eugene was 16 years old, when Jeantel was 19 (Jeantel said that she lied about her age for some strange reason).
As Huemer says:
All of this apparently sounded perfectly legitimate to the journalists who covered the case at the time. If it sounds completely reasonable to you as well, I have several bridges I would like to sell you.
If this is right, then it seems clear enough that Zimmerman was acting in self defense or at least was not provably guilty of murder. It’s still possible that he embellished details. Here are three alternative possibilities:
Zimmerman approached Martin telling him to stop. Martin attacks Zimmerman and gets on top of him. Martin is beating the shit out of Zimmerman but doesn’t grab his gun. Zimmerman shoots him. In this case, it would probably still be self defense, but much more debatable.
Zimmerman approached and attacked Martin. Martin is stronger and gets on top of Zimmerman and begins beating him up. Zimmerman shoots him. In this case, as Zimmerman was the aggressor, he would be guilty of murder.
Maybe neither was the clear aggressor. Perhaps Zimmerman went up to Martin and told him to stop and wait for the cops. Martin then shoved him, Zimmerman tried to restrain him, and the two ended up in a fight. Martin ended on top of Zimmerman, beating him up, and Zimmerman shot him.
All of these—especially the third—seem decently plausible, certainly not much less implausible than Martin just randomly attacking Zimmerman. Still, in the absence of any clear evidence for them, there’s obviously not enough to convict. For that reason, I disagree with Huemer that Martin obviously attacked Zimmerman, but think that Zimmerman was probably on the ground, getting beat up, and therefore acting in self-defense. Certainly if Huemer is right, there isn’t enough evidence to convict. So the question is simply: is Huemer correct in his portrayal?
3.2 The facts
The claims about Jeantel look like they hold up decently well. The case seems sufficiently overwhelming and Jeantel’s claims are so unbelievably implausible that it looks like she was completely full of shit.
The one thing that Huemer left out but isn’t important is that several different neighbors reported hearing the goings-on but most of their testimony isn’t important. Selene Bahadoor said she couldn’t see who was on top as it was too dark. Jayne Surdyka said she heard someone cry for help who she thought was “the boy” but described being unable to be confident. Jeannee Manalo reported being unable to see who was on top until she saw the news, at which point she thought it was Zimmerman—but this is a classic example of an unreliable witness, whose memory of an event is affected deleteriously by their judgments after the event. The only neighbor who had some specific memory before seeing the news about it is Jonathan Good, the witness Huemer mentioned who corroborated Zimmerman’s story.
Other than failing to note the presence of the other witnesses, what Huemer says is basically right. In fact, many friends of Zimmerman testified that the screaming in the background of the 911 call sounded like his screaming. So the case is perhaps even better than he thought.
My conclusion: Huemer’s claims hold up. There’s certainly not enough evidence to convict. It’s not terribly unlikely that both were at fault to some degree—especially given Zimmerman’s subsequent crimes—but probably Zimmerman was on the ground, getting severely beaten when he shot Trayvon.
4 Overall takeaway
It’s rare that a book actually changes your mind about some significant subject. Huemer’s book is an exception. While I remain staunchly anti-Trump (as is Huemer) and a liberal, this has made me more economically conservative. Regulations do seem plausibly extraordinarily bad.
There are things that progressives say. You hear them a lot if you hang out in the spaces that I do. While I’m already pretty skeptical about them, this book has made me more skeptical. If you learn that many progressive claims are built on sand, you should be more skeptical about the other ones that you haven’t investigated. Many of the memes floating around about politics on both sides are total nonsense.
There’s a nearby possible world where a book like this would make me a Republican. If the Republicans were still pro-PEPFAR and had Romney’s views, for instance, and the Democrats supported tariffs and price controls, I could easily see myself voting for the Republican. Unfortunately, as long as the Republicans keep picking a candidate who makes concerted efforts to overturn fair elections, I will remain a firm Democrat.
I am an economist and I regard that paper as rather weak evidence for the case that regulations reduce growth, for the same reason alluded to in the post: all the variation is over time in a single country.
From what I can tell, the paper was published because their time series analysis is integrated with a macroeconomic model such that they can parse out a relationship with total factor productivity, and because their results are in line with other papers that don’t rely solely on time series analysis. As always, one should draw inferences on the basis of the literature as a whole and update only weakly on individual papers. If the cited papers are using cross-country analysis this is better than time series analysis but has its own issues, and it makes some sense to see what happens if you try a pure time series approach, which the authors of this paper did. Ultimately it’s difficult for us to know much about a causal effect of total regulation because it is composed of many heterogeneous regulations and jurisdictions are large and few enough that we can’t really apply the sort of quasi-experimental methods that are the standard for estimating causal relationships.
The bull case for deregulation comes from looking at all the individual cases where overregulation is deeply harmful (housing, project permitting, drug development, nuclear, occupational licensing, etc.) and becoming redpilled to the apparent tendency of progressives and the federal bureaucracy to impose grievous costs for scant gains (though obviously some regulations are necessary). But it’s not enough to make some regulation that under ideal conditions results in some good thing and then pat yourself on the back for a job well done, there are usually assumptions that don’t hold, unmeasured costs, and unintended consequences.
(Edited for paragraph spacing)
A technical point: it’s not really accurate to say regression analyses are regarded as good evidence. It heavily depends on why and how you conduct the analysis.
It’s like saying “using microscopes is considered good evidence.” Well… sometimes?? Depends what you’re looking at and what you’re concluding based on that. If you’re trying to find the weight of an elephant, using a microscope is the wrong tool. Regression analyses can also be the wrong tool in many circumstances.