Economics Is a Science, Prelude

Against political epistemic nihilism

Introduction

In college, when I told people I was doing a Bachelor of Science in Economics, the responses included “My boyfriend likes to call it astrology for men”, “Astrology for men!”, “Bullshit economics!”, and other things like that. They were overwhelmingly progressive Democrats. As I write this, the American economy is struggling under the weight of tariffs imposed by President Trump. Both large financial institutions like JPMorgan Chase and betting markets believe a recession is likely this year. The political persuasions of the people who say and do things that ignore the advice of economists differ greatly, but their doubts are the same.

You may doubt that the way we think about economic issues can be objective. In reply to one of my many blog posts, a commenter asserted that there is no correct view about the economy, regardless of how many academic papers you cite. This series of posts argues they were wrong.

But it’s notable that you really can almost always find evidence in favor of your position when talking about economics. The problem with this view is that the same criticisms can be wielded against other sciences. You can find evidence that the minimum wage causes unemployment, and evidence that it does not. You can also find evidence that homeopathic medicine—widely decried by doctors and described as a pseudoscience on Wikipedia—works.¹ Economies are complex, and so is the human body, and reality in general. That doesn’t mean you should give up and believe whatever you would like to.

I’d like to show you that economics adheres to the standards of modern science. In truth, there’s no single scientific method. But there are common features present in various scientific disciplines and not present in what we usually call pseudoscience. To see just how scientific economics is, we can look at many of those features and decide whether the field of economics has them. This is not meant to show you that you should believe everything economists say, but to show you where your doubts should lie. Most people already know that it’s wrong to dogmatically believe whatever experts tell you. It’s often more important to say that doubting everything is just another form of ignorance.

People who doubt everything are easy to find, since the tendencies of mass media to mislead and confuse people are easy to see. During the pandemic, a skeptic might have said “The vaccinations, the masks, the lockdowns—they just want to control people.” Someone who strictly adheres to what seems to the advice of scientists might have said “You’re nuts. Get your kids vaccinated, wear a mask, and follow the rules.” It’s easy to see that the second opinion here is much safer. But our dogmatist would also have had a hard time seeing the long-term impact lockdowns had on the development of children.² Meanwhile, the skeptic was more likely to die of COVID-19.³

I’ve met people who perceive economists to be dirty capitalists who wish harm on the working class, and oppose things like unions for that reason. That’s not what the work of economists looks like. Economists often oppose unions because labor markets appear competitive, meaning unions can only boost wages by excluding workers who might compete and drive down wages. But the real work economists do involves creating models of union behavior and then testing those models against reality. In truth, some economists are friendly to unions because they know they can help under certain circumstances, though none of them would say they’re always good. To add another example, a student of mine was disgruntled by the mention of economist Milton Friedman in class, given Friedman’s pop-culture appearance of unbridled libertarianism. I tried to persuade him to think differently about Friedman by pointing out how he wanted a negative income tax, a form of universal basic income, and opposed the cartel among wealthy doctors. But his prior belief about him, like most beliefs about economists, seemed stuck.

Economics is hard work, and on closer examination, the opinions of economists aren’t easy to predict without understanding economics itself. You would do very poorly by assuming economists always support the status quo of capitalism. Many will oppose unions while encouraging wage subsidies in the same breath, rather than a completely hands-off approach.⁴

Note that sources in this series are usually only provided when they aren’t easy to find with Google (and yes, I check). Often people ask for sources when they disagree with a claim, and then refuse to go looking for them. This makes sense, because one shouldn’t be expected to prove themselves wrong when it’s in the interest of someone else to do it. But if everyone involved is interested in learning, rather than having a disagreement, this is not necessary. So I’ve avoided making the text appear too loaded with underlines and footnotes by being sparing with citations. Finally, much of this book was written with the help of Theory and Reality: an introduction to the philosophy of science, written by Peter Godfrey-Smith and published in 2003. This provided some additional, useful information about what philosophers think counts as “scientific”, and should be assumed to be the source when reference is made to philosophers of science and their ideas.

Against Political Epistemic Nihilism

Epistemic nihilism refers to the idea that nothing can really be known about the universe. Political epistemic nihilism is a wordy but descriptive way of referring to a kind of thinking that has become pervasive in American politics. Here is roughly what this looks like:

The world is a mess, rife with elites and their journalists who are happy to lie and mislead, constantly. They are destroying this country from the inside out. If we don’t do something, they will. And they’ll tell you that I’m a liar, a thief, and a cheat—and hell, sometimes they’ll be right!—but know that for everything they say about me, they’ve done worse. They’re just better at covering it up.

This is the essence of Donald Trump’s epistemology (i.e., his way of deciding what is true), if he can be said to have one. His ideas are appealing to people who have lost faith in American institutions. If you believe that American journalists, politicians, and other elites are all liars, then it is appealing to put your complete faith in someone else, like Trump, who might be more trustworthy.

The element of faith is key. What is unique about political epistemic nihilism is that it inevitably leads to an embrace of one’s moral intuitions and faith in a preferred leader, rather than complete doubt in knowledge or a desperate search for some new way of understanding the world. So the name I’m using is somewhat misleading—this kind of epistemic nihilist has very passionate and fixed beliefs, but this is precisely because they don’t have enough trust to learn much about the world.

This kind of thinking is not exclusive to Trump and his many followers. Two-thirds of Americans have “not very much” or no trust at all in mass media like newspapers, TV, and radio to report the news fully, accurately, and fairly.⁵ References to Noam Chomsky’s Manufacturing Consent are common on the political left. But there’s no drought of conviction, and no scarcity of action, among anyone engaged with politics.

I’m addressing this way of thinking because if you believe that elites are happy to make things up, no description of the field of economics will convince you that it is accurate. I will take a moment to clarify exactly why I would expect that. This is not a description of how information works in practice, but it makes the problem clear.

Suppose that you are arguing with a friend over whether President Brugno IV sexually assaulted a journalist named Carly Kristen. You ardently believe that he did, and you even provide a source saying that he did. But your friend doesn’t believe that the source is accurately reporting information. They go as far to say that they made the story up. How do you show that they didn’t make it up?

The common response in this situation is to provide even more sources. You might even provide sources that you would assume they would trust, like Box News. But that runs into the same problem: you can always just say that it was made up. Searching for advice on how to identify credible or reliable sources doesn’t really provide much help either. I found an article on a university’s website that tries to help people find credible sources, and it tells you to examine the authority of the source, compare it to other sources, and look for citations within it. But even if the source is an expert, it matches other sources, and the source cites academic papers, none of those things will really help you. (The responses from your skeptical friend could be “Experts lie, the other sources are making it up too, and anyone can come up with bullshit citations. Try again!”)⁶

The main point of this article is that political epistemic nihilism is bad, but rational arguments are not in themselves sufficient to show that it’s wrong. There are many arguments that can be made to dissuade you from it, but as you’ll see, none of them are like a deductive or mathematical argument in how purely correct they are. I would hope that the remainder of the section would somehow prove convincing, but even a purely rational and calculating political epistemic nihilist could survive it and believe, for example, that Donald Trump would have won the 2020 presidential election if not for widespread voter fraud.⁷

Chained Reality

We can proceed through another example, taken from Scott Alexander’s blog. If you’re a Republican or you just happened to vote for Kari Lake in Arizona, you’d be very concerned about voter fraud occurring that favored Democrat Katie Hobbs. So it’s concerning when you see a headline like “Kari Lake Trial Bombshell: Audit Reveals 42.5% of Ballots Randomly Sampled Were ILLEGAL Ballots”, as run by InfoWars.⁸

Scott clears up what’s really going on in this case. He shows how the government explained they might have printed ballots with mistaken sizes, but they would have been counted by hand anyway, so there’s no issue. Then he says that InfoWars was not really lying—the problem was that they were focusing on the Lake campaign’s claims while ignoring everything else.

Returning to the main issue, how does an observer know that any of the evidence provided is truthful, anyway? The Lake campaign could have made it up. The government acknowledged the issue and explained it, so it seems like they didn’t make it up. But the government could have made that up, too. This is the same problem as before.

Here is one solution. In the story, “the government” refers to the Assistant Maricopa County Attorney, Tom Liddy, who explained the issue. If you assume that the government of Arizona really does exist, then this report must really come from Liddy. If it didn’t, either he or someone in contact with him would have come out with the news that this is made up, or someone could reach out to his office and find out. If he was lying about the way ballots are counted, again, someone else within the government would feel compelled to say something. It would easily catch the attention of a journalist. Even an out-of-state lawyer might know about the ballot-counting procedure and see the article.

You might counter that everyone in the government of Arizona was in on it and would never have come out against Liddy. This seems to require assuming that Republican officials and those sympathetic to them, who controlled the legislature at the time, were also in on the conspiracy, which appears plainly false. More importantly, it requires that everyone familiar with how ballots are counted in Arizona did not speak up, and that even includes people outside of the state, many of whom would be motivated to call out the lie.

I’ll write something absurd here to clarify my point. Say you doubt the existence of the government of Arizona in the first place. After all, not everyone has been to Phoenix, and not everyone in Phoenix has even visited the Capitol building. But if that building didn’t exist, people in Phoenix would quickly notice. If Phoenix didn’t exist, people in areas bordering the city would notice (or maybe astronauts on the ISS). I would have noticed, since I’ve been there.

It’s nice to be able to verify claims, and this kind of almost-verification I’ve described is made possible by the things you can directly observe. Wouldn’t you notice if a building in your city that people talk about doesn’t exist? Wouldn’t you at least hear about that from a friend or colleague at some point? The things I’m saying you should expect from other people are also things you can observe in your own life. The entire planet is bound together in this way, almost guaranteeing the existence of places like Antarctica and New Zealand and people like Tom Liddy. You haven’t been there, and you don’t know the people there, but the world is chained together such that it doesn’t matter. This is a valuable observation to make. It gives us great confidence that the things the media reports on are really true: if they were not true, we would expect many other things to eventually be very different, in particular the existence of contrary claims.

So, political epistemic nihilism often requires making absurd assumptions about the way reality chains together. We would expect a lot of things to be different if Liddy wasn’t reporting accurately on the ballot counting procedure, in particular, we would expect other people familiar with the matter to contradict what he’s saying. Reports about any distant environment (e.g. government operations in Phoenix) are likely to be accurate when sources are connected to that environment and that environment is connected to you.⁹ We’ll also have to add as a condition that there isn’t contrary testimony from observers in the same environment.

Consider the case of the Springfield pet-eating hoax. This is a somewhat difficult case to deal with, because the supposed incidents were sporadic rather than constant, unlike the presence of a building, person, or process, like how ballots are counted. But if Haitians in Springfield, Ohio really did eat pets and this was witnessed multiple times, you would expect such witnesses to come forward about it, or for photos of butchered cats to appear. Neither of those things happened. The things you would expect to happen if pets were being eaten in your own town didn’t happen in Springfield. No Haitians were arrested and convicted of pet eating.

The original Facebook post that sparked the story instead described rumors rather than first-hand experience, and the author, Erika Lee, later deleted the post and expressed regret about making the post to NBC News.¹⁰ That, or NBC News made it up, Lee’s Facebook account was seized by someone else, and she has not been able to be reached for comment, which sounds implausible. If NBC News made it up, Lee could have posted about it. If she was somehow cut off from the internet completely or kidnapped, it would have been huge news among right-wing media outlets, who could easily find that she went missing. I could go on, but the point should be clear.

The Springfield News-Sun, a local newspaper, reported that the local police department said they were aware of the social media post that sparked the story, but said this was “not something that’s on our radar right now.”¹¹ This local paper must exist, as well as the police department, so this reporting should be accurate.¹² If people were eating pets and plausible evidence existed, we would expect the police to have caught wind, but they didn’t. This way of thinking might not erase all doubts, but it’s pretty good.

I have perhaps been too harsh on the American right, so let’s turn to an example from the American left. This is not meant to be representative of the way Democrats or Democratic Socialists think, nor should previous examples be taken that way in regards to fans of the president (mostly because I don’t want to add anything to this piece I don’t need to defend to make my point).

I came across a post on X, formerly Twitter, claiming that Cubans live longer than Americans.¹³ This didn’t sound right to me, so I decided to check, and indeed, the opposite is true: Americans have generally been slightly longer-lived than Cubans.¹⁴ I shared this fact in a reply.¹⁵

This led to an interesting exchange with another user. The user claimed that Cuba’s life expectancy is now longer than that of the US—the source I had provided from the World Bank was cut off at 2022—but didn’t provide a source. I provided more sources from the Pan-American Health Organization and the Congressional Budgetary Office to show they were wrong, and that American life expectancy is just slightly higher than Cuban life expectancy, by 0.5 years.

They then told me that was not the point, and that the point is that Cuba, a poor country, is taking care of its population while the US, a rich country, uses its population. (Their original wording wasn’t much more clear.) I then asked for evidence that the typical Cuban enjoys either a longer life expectancy or a better standard of living than the typical American, and they responded by saying that my own evidence was wrong because it’s propaganda from the United States federal government. (You can see where this is going.) I told them they should trust the evidence I gave, because it’s unlikely that American propaganda would tell you Americans are shorter-lived than Europeans. They again claimed it’s propaganda by saying Congress is “literally owned by israhell”, referring to Israel. I gave them another source, and they rejected it because it was partially funded by Michael Bloomberg, a “billionaire who peddles propaganda”. I gave them another source, Macrotrends, but admitted with regret that Macrotrends did not say where their data was coming from. Macrotrends actually showed that Americans live slightly shorter lives than Cubans, but are still on par (as all the sources showed). They responded with the following:

They blocked me after giving this reply. I never had any hope of changing their mind.¹⁶ The very problem that seemed to be causing them to believe Cubans live longer than Americans, confirmation bias, was exactly what they accused me of, despite providing every source I could find. They were a political epistemic nihilist: anything that felt wrong to them was automatically suspect, they did not care about providing their own sources, and they gladly believed whatever they wanted to.

This exchange tells you a lot about the problem I’m trying to confront here. The idea they have is completely implausible on close examination. Consider just their rejection of the CBO’s data, based on the claim that Israel owns Congress. The CBO is a government agency, so what they’re really claiming is that Congress appointed nominees to the CBO who would aim to inflate American life expectancy such that it happens to be slightly higher than that of Cuba.¹⁷ Not only is it hard to conceive of a motivation to do this, especially if this is at the direction of Israel, but there would also be clear signs that this had happened from American journalists, who are not exactly friendly to the government and write about the nominations made to government agencies, especially when notable.¹⁸

Let’s steelman their argument and consider a more plausible claim, that those at the CBO are biased in favor of their own country and fabricated data to make recent life expectancy statistics seem higher to make up for a remarkable deficit. Because of how life expectancy is calculated, that would mean that they either raised the ages at death of people who recently died or published a number with no relation to their data. It would also mean that someone could show they’re lying by simply performing a randomized national survey of people who know someone who recently died. All of this would, of course, put the reputation of the CBO at risk, to create a statistic that does not directly benefit anyone involved.

The overarching point here has been that even though we often can’t directly examine reality to decide if a source is accurate, there are good forms of theoretical reasoning we can use to decide. In particular, we can notice what would happen if we assume something is a fabrication. Every source exists in the real world. If it was a fabrication, it would be difficult to maintain secrecy about the conspiracy. Each person involved in the conspiracy is like an extra dice roll that might lead to a leak, and players of any board game know you’ll roll snake eyes eventually.

By this point, we have fairly good reason to believe the data and other information people provide under certain circumstances. When fabrications occur, we expect other people familiar with the matter to notice, and when they don’t, it seems unlikely that something was made up. There are also some simpler heuristics I brought in, like how it seems unlikely for an organization like the CBO to make something up when they have no clear motivation for it.

But there are still problems. A professional skeptic would still believe it’s possible that the CBO successfully eliminated a larger deficit in American life expectancy by fudging its numbers, and that the conspiracy was successful because of strong norms within the organization and the government as a whole. Nobody produced a contradictory survey because such a survey can only be performed with the resources of a large organization, and any such organization would quickly be infiltrated by the government.

Moving on, we’ll now see how contrary testimony makes the task even harder when it does exist.

Contrary testimony, conspiracy, and other arguments

Cursory searching for evidence supporting the pet-eating hoax did not lead me to any. But it did lead me to J.D. Vance’s claim that such first-hand testimony exists.¹⁹ I would like it if there were some way to show that he was definitely lying, but there is none.

You could, for example, point to how much more evidence there is that the pet-eating hoax was made up. This seems like a good way to think about the world, and I would expect a jury to acquit someone if there is simply more evidence in a defendant’s favor than against. But it doesn’t provide complete certainty, because there is still the possibility that the police were incompetent and news organizations conspired to hide evidence. You could leap into extreme schizo land and talk about how the people who witnessed the pet eating were assassinated, and we would only be able to say that what you’re saying is highly unlikely, not impossible.

Alternatively, you could point out how Vance failed to point to exactly who this first-hand testimony is coming from. This is similar to how the fan of Cuba I was arguing with didn’t provide sources of their own, because saying there’s evidence is not the same thing as providing it. But here, one might simply trust that Vance had this evidence, and there’s nothing logically incorrect about that, even if it sounds unlikely to most observers.

There are further arguments one could try to use to beat out political epistemic nihilism, aside from the contradictions introduced by assuming something isn’t true and the increasing difficulty of maintaining a conspiracy as it expands. Let’s return to the example of the CBO and run through them.

In the philosophy of science, some philosophers believe that Occam’s razor should be applied to scientific explanations. This tells you that in general, simpler explanations should be preferred, and more complicated explanations should require greater evidence. But as described by philosopher Peter Godfrey-Smith, “Despite various elaborate attempts, I do not think we have made much progress on understanding the operation of, or justification for, this preference” for simpler explanations.²⁰

Perhaps he just needed to wait, since he was writing in the early 2000s and further evidence has arrived. In a 2015 review of evidence for the Journal of Business Research, Green and Armstrong write that “Complexity increases forecast error by 27 percent on average in the 25 papers with quantitative comparisons.”²¹ In practice, it seems that Occam’s razor can be good, at least for some applications.

But there is no good reason to expect it to surely be true in general. If you applied Occam’s razor to the CBO example and preferred to believe the data is real because it’s simpler than supposing there is a conspiracy driven by Israeli ownership of Congress, you would still not be able to clear away all doubts. Nothing can completely defeat the idea that there is a sufficiently deceptive and effective conspiracy to trick us into thinking Americans live about as long as Cubans, rather than much shorter lives.²²

Falsifiability might be used via an argument along the lines of “If there is no way to falsify your belief in this conspiracy, your belief is not scientific. Clearly there is not, so your belief isn’t scientific.” But this belief really could be falsified. If they were personally able to help a communist organization’s randomized survey of Americans to acquire a life expectancy statistic, they could falsify it. Worse, even if the conspiracy was unfalsifiable, they have no obligation to believe that falsifiability decides whether something is scientific.

The Cathedral

Next, we should look at a particular breed of scientific skepticism so that you might better appreciate the ease with which a sufficiently good thinker can escape belief in science. This comes from Curtis Yarvin, a conservative blogger who is notable for some unusual beliefs like his belief that an accountable monarchy would be superior to American democracy. He has apparently been influential as a blogger, or at least has found some importance, since VP Vance recently decided to follow him on Twitter/X.

Interestingly, his essay “A brief explanation of the Cathedral” implicitly acknowledges the problem I’m confronting here, that being the problem that there are only theoretical reasons to believe distant sources, since we cannot directly observe things we care about in politics. The same is true in most sciences, which is part of what Yarvin is writing about here with his general theoretical reasons to distrust science and journalism.

Yarvin has a habit of writing about these issues poetically, referring to journalism and academia as “the cathedral”.²³ His point in using this term is that while the Catholic Church was one institution, the cathedral was many, and central to his ideas about how journalism and academia work is that the many cathedrals of journalism and academia have become one. His explanation of why also tells you why you might not trust them.

Yarvin presents one form of the “marketplace of ideas” approach to science. Karl Popper, who famously argued falsifiability should decide what counts as science, was one philosopher who believed in something like this. To Popper, some scientific ideas are falsified while others stand up to testing and are naturally selected, like animals, to survive and spread. The new idea Yarvin introduces is that journalism and academia do not select ideas for their ability to stand up to testing (except in the physical sciences, which he is not focused on), but select them for their ability to validate the use of power. Climate alarmism, for example, is such a “dominant” idea, as Yarvin terms it, because it tells you to spend lots of money on fixing the problem. He believes that he and other more conservative thinkers are free from this influence and, operating underground, are precisely the rebels you should trust to accurately characterize reality.

Yarvin’s essay does not make sense. The most glaring issue I found was that he named “tolerating crime” as a dominant idea. “We should tolerate crime” tells you to not exercise power. He only included this because liberals have often been attacked for allying themselves with progressives who want to cut back spending on policing. In general, it doesn’t appear that journalists and academics select for ideas that justify the use of power. They also, for example, tend to be pro-choice, which is one form of opposition to the use of government force.²⁴

The stronger idea is this: if academics and journalists are connected to politicians, ideas that justify the use of power might be favored. Observing this connection should make you turn your nose at the evidence they provide. This is a decent theoretical reason to avoid the Cathedral, as Yarvin calls it. Yarvin’s thought is even more expansive on this issue. He argues that these sources cannot be trusted to, for example, report accurately on genetic differences between white people and black people, since many of them openly declare that defeating racism is one of their goals. This is luckily more of a problem for sociologists than economists, who generally do not fixate on racism, though they often study it (e.g. Bertrand and Mullainathan 2003, or a replication that also found discrimination against black applicants in 2024).²⁵

What Yarvin is offering is still different from what most political epistemic nihilists, and most Americans, are choosing. Their skepticism has often turned them towards random people on social media instead. This does not look necessarily bad to me, because I know there are good thinkers even on TikTok and X/Twitter, like Christopher Clarke, an economist.²⁶ But it does mean that in practice, what the marketplace of ideas most people get their information from selects for is the ability to get attention. I suspect this often means two things: making people feel angry, and reaffirming ideas we are emotionally attached to. During the economic expansion that followed COVID, real wages grew most quickly for the lowest earners and inequality declined.²⁷ But evidence showing this is true could never be selected for by social media algorithms or preferred by an American audience.²⁸ That’s why consumer sentiment was so low during the rapid growth in incomes following COVID.²⁹

Secure Communities and Fabrication

For the purposes of the series, we care more about the data used in econometric work, like the work examining the effect of deportations on crime. So let’s take a moment to look at the econometric work around the Secure Communities program. This program increased the rate of deportation, and was deployed earlier in some places than others based on their preparedness, making increases in the deportation rate associated with the program essentially random. Places where such increases occurred did not see decreases in crime distinguishable from zero.

Let’s focus on the crime data they use from the FBI’s Uniform Crime Reporting. This data is not collected by the FBI, but by local law enforcement agencies, who then report crimes to the FBI. To make sure the evidence here is good, we only need to know two things: that the data is being collected from the real world, and that reports of crime are unrelated to the Secure Communities program. You might claim that undocumented immigrants are less likely to be caught, making the analysis invalid. But even if that were true, what matters most is variation in reported criminal activity, not the absolute level. We would still expect to see the crime rate fall in the counties where Secure Communities was deployed earlier, compared to others, especially given the size of the dataset.

Whether the data is collected from the real world or not, or has been subject to some kind of fabrication, does not seem obvious. How could you know, except by creating your own network of organizations for data collection on crime? Return to an earlier idea about conspiracies. It’s easy to lie, but it’s very difficult to get hundreds of people to all tell a lie without anyone revealing evidence of the scheme. So the decentralization of the FBI’s crime reporting helps a lot. To successfully conspire to rig data on crime to make Secure Communities seem ineffective, you would have to deal with thousands of reporting police agencies across the United States. The FBI could rig their reports directly by working with every police department to do so, which would inevitably lead to a leak. The FBI could modify the reports upon receiving them, and hope none of the police agencies check to make sure their data is reflected accurately in the database, which also sounds unlikely given the great risks. Finally, economists themselves could modify the data, and hope that nobody doing peer review, no statisticians looking at their work, and nobody in the FBI itself or any of those police agencies checks to make sure the data is fine.

There seems to be another problem: Secure Communities began earlier in counties that were better prepared to cooperate with the federal government. Couldn’t it be that they only reported more crimes, compared to what we’d expect if deportations reduced the crime rate, because they were better at finding them? But for these counties to counteract the effects of deportations on crimes, we would need them to report more crimes once the program began, rather than simply reporting a higher absolute level of crime, which is what we would expect based on their preparedness. So the data appears trustworthy, and what matters is whether the analysis is valid. That’s a question you’ll only be able to answer later in the post on econometrics.

Conclusion

We’ve seen a variety of arguments meant to encourage belief in evidence concerning issues of importance to economics, politics, and culture:

• Reality chains together, so assuming something is made up often leads to strange conclusions. If a bombing in Gaza had not really occurred, it would be easy for the many thousands of people living there with internet access to report on that fact (though most there don’t have access). If China was not polluting as much as reported, that would be reflected in measures of global emissions, too.

• Conspiracies become difficult to maintain as the number of people involved rises. To give an example, researchers found that vaccines are not related to autism, and nor is thimerosal, a mercury-containing compound, or mercury itself. If evidence is all you need, this meta-analysis should be enough to sway you. But you might suspect that doctors simply wish to continue selling vaccines. If the data is made up and their observations didn’t really happen, there are massive opportunities for someone to leak damning evidence of fraud. Any of the researchers could have created that leak, and there are many, since this is a meta-analysis. Strange patterns in the data might have been spotted in peer review, too.

• Sometimes, it isn’t clear why someone, or an organization, would be motivated to lie about something. It would be easy to get people to believe life expectancy in the US is comparable to that in Europe, but it’s closer to that of Cuba, despite the potentially-biased CBO reporting on it.

• People with a clear motivation to contradict the claim in question might choose not to. They may even support the claim instead, like how the Heritage Foundation gladly publicized counts of voter fraud too small to swing the 2020 presidential election.³⁰

• Conspiracies sometimes appear to be unfalsifiable and thus, by that standard, unscientific.

These are useful ideas to shield yourself from nonsense. But they aren’t decisive, and will often only be useful to people who already believe in empiricism, journalism, and academia. A sufficiently strong conspiracy would still be able to hide the truth from you. A lack of a clear motivation for a conspiracy does not guarantee that it doesn’t exist, since people are not always rational. And even if something appears unfalsifiable, nobody has an obligation to adhere to Karl Popper’s ideas. The reasoning that allows you to escape these arguments sounds silly, but that doesn’t mean it’s wrong.

I’m not writing this to try to kill any hope of convincing people they’re wrong. The point is that there is no argument I can give that will guarantee your belief that, for example, forcing landlords to charge less rent reduces housing supply. Some kind of empiricism is the foundation to all sciences, but unless you show some very basic generosity and faith in it, science doesn’t let you know anything.

So while I oppose political epistemic nihilism, I’m not aiming to rebut that brand of thinking. If you are invulnerable to evidence, you’ll continue to support things like tariffs no matter what arguments are presented in favor of economics as a science, because the arguments that can be deployed against it can be deployed against other sciences as well, like pharmacology.

The next post will be much longer, and will compare economics to other fields of science. It will not be very useful to you if you’re a nihilist about information, or if you apply a much higher standard to social sciences than you do to others.

1. https://slatestarcodex.com/2014/12/12/beware-the-man-of-one-study/

2. https://pmc.ncbi.nlm.nih.gov/articles/PMC10266495/

3. https://pubmed.ncbi.nlm.nih.gov/37486680/

4. The use of the term laissez-faire always seemed pretentious to me. It adds additional needed instruction without improving on the depth or clarity of the field. Teaching its meaning is only necessary because of the overuse of the term by academic economists.

5. https://news.gallup.com/poll/651977/americans-trust-media-remains-trend-low.aspx

6. https://www.stevenson.edu/online/about-us/news/how-to-identify-reliable-information/

7. If you care about evidence, you can check out this article from Brookings, which itself cites the Heritage Foundation. Somehow, Heritage both authored Project 2025 and authored evidence showing that the 2020 election was not rigged: https://www.brookings.edu/articles/widespread-election-fraud-claims-by-republicans-dont-match-the-evidence/

8. The Media Very Rarely Lies

9. Here, “connected” means “capable of being truthfully recognized in some way, either directly or through an intermediating link”.

10. https://www.nbcnews.com/news/us-news/-just-exploded-springfield-woman-says-never-meant-spark-rumors-haitian-rcna171099

11. https://www.springfieldnewssun.com/news/springfield-police-say-no-reports-of-pets-stolen-after-viral-social-media-post/3WSIZQNHQVE4NP4TS5BVHBB2PY/

12. If they did not exist, it would be fairly likely that at least one person in Springfield, a town of over 50,000 people, would notice.

13. https://x.com/mcsquared34/status/1911744390441193773

14. https://data.worldbank.org/indicator/SP.DYN.LE00.IN?locations=CU-US

15. Not really to change minds, and more as practice to cement important information in my own mind.

16. (I knew that.)

17. The use of “appointed” here is not a mistake. Unlike the directors of other agencies, the Speaker of the House of Representatives and the President pro tempore of the Senate jointly appoint the CBO director. https://www.congress.gov/crs-product/RL31880

18. e.g. the appointment of RFK Jr. to head the Department of Health and Human Services. He was a well-known antivaxxer, but thankfully, he has at least changed his tune on the MMR vaccine. https://www.npr.org/sections/shots-health-news/2025/04/07/nx-s1-5354900/hhs-rfk-endorses-mmr-measles-vaccine-stoking-supporters-fury

He also has rejected the germ theory of disease, believing that illness is primarily caused by poor diet and environmental toxins.

19. https://www.bbc.com/news/articles/cgj447j5711o

20. From Theory and Reality, page 214.

21. https://repository.upenn.edu/entities/publication/d3ca99d6-2990-435f-a2fe-37d7d325323e

22. The absurdity of the belief I’m talking about here doesn’t escape me; sometimes it’s worthwhile to take an absurd idea seriously to improve our own understanding.

23. Much to my displeasure, since I can only imagine much of his audience is reading his work and taking him seriously because he sounds smart. In this case, he gets a pass because he explains his purpose in calling it the cathedral.

24. Libertarians make exceptions to their opposition to government, such as by supporting a smaller state that would maintain order and criminalize murder. So while it takes some extra reasoning to integrate the pro-life position into libertarianism, it’s not complicated. I’m more surprised by a user I found who thinks that immigration restriction is obviously libertarian and requires no justification.

25. https://bfi.uchicago.edu/wp-content/uploads/2024/04/A-Discrimination-Report-Card-1.pdf

26. Do assistant professors count as economists? I don’t know, and I don’t care that much. https://x.com/EconChrisClarke ; https://www.tiktok.com/@econchrisclarke

27. https://www.nber.org/system/files/working_papers/w31010/w31010.pdf

28. It seems obvious that optimistic information is generally slower to spread and does not drive engagement.

29. https://fred.stlouisfed.org/series/UMCSENT

30. Do note that Heritage is highly unreliable. Here is a page on their website that lies about the electoral college. They claim that the electoral college helps avoid contentious election results by magnifying the margin of victory. This is incorrect. The electoral college relies on smaller, state-level popular votes to function, meaning that elections decided by a small number of votes are more likely, not less. A national popular vote, which involves a much larger population, naturally has larger margins even when the distribution of voters between parties is the same. In practice, the smallest national popular vote margin was 1,898 votes in 1880 between James A. Garfield and Winfield Scott Hancock, while the smallest electoral college margin was 537 votes in Florida between George W. Bush and Al Gore. (The latter election took place at a time where the total population of the US was much larger, too.)