Against Elegant Political Models

And why I'm riding with Biden

1

Vaishnav Sunil

—substacker, manifest attendee, and friend—has an article called The Blind Centrist's Guide to Gaza. It begins with what is, I think, a pretty good criticism of centrists—that often, on all sorts of political issues, the correct view is not precisely in the middle. On the issue of meat-eating, for instance, and immigration, the truth is not in the center but is instead quite extreme, so extreme to be so thoroughly outside of the Overton window. Centrism is a poor heuristic that goes wrong in lost of cases.

In his piece, Vaish argues from something like first principles that a response like the one Israel is carrying out is justified. He argues that this can be deduced from the incentives on both sides—Hamas has an incentive to repeat the act unless Israel shows vastly disproportionate force. One needn’t look so much at the details of the war: from first principles—or something in the vicinity of first principles—one can figure out that the war is justified.

While Vaish is great, I think this is badly mistaken. Figuring out if the war is justified requires actually looking at the numbers and figuring out what obligations a nation has to its own citizens vs citizens of another country. It can’t be figured out merely by constructing an elegant model of incentives.

Imagine that the only way to eradicate Hamas would be to drop a nuclear warhead on Gaza. While this would be wonderfully effective in terms of incentives—Iran would think twice before messing with Israel and Hamas would be no more—it would be seriously wrong. The fact that some action gives one’s enemies good incentives counts in favor of it, but isn’t enough to guarantee that the action is worth taking.

Similarly, imagine that a war like the kind Israel is launching now would require killing 500,000 people. Surely then it wouldn’t be justified. Again, such a war would give Israel’s enemies the right sorts of incentives but would still be deeply wrong. For this reason, figuring out if Israel is in the right requires a lot more than armchair pontificating—it requires forming complicated judgments about how many people Israel should be allowed to kill in the pursuit of Hamas’s eradication, how weakened Hamas will be by the end of the war, and how many people will be killed.

Certainly, such a priori reflection can tell us some things about the war. But it can’t tell us enough to figure out if the war is justified.

This is true of all sorts of armchair political science speculation. The real world is a messy and complicated place, and while models can tell us some things, they can rarely settle issues. For this reason, unless one has carefully combed through the facts surrounding some topic, they shouldn’t be confident in their views on that topic.

When I was a libertarian in middle school, I thought politics was mostly an armchair discipline. I thought that most of the reason people disagreed with me was that they were basically ignorant of economics, that they didn’t understand how economic incentives worked. For example, I reasoned that government regulations from entities like the FDA were unjustified because, if someone sells a bad product, they could be sued—and people won’t go back. Similarly, I thought the minimum wage had to be a bad idea—basic economics says that it increases unemployment.

As I got older, however, I realized these elegant models were oversimplified. Whether producing a dangerous product will be profitable depends on facts about the world: how expensive it is to produce a dangerous product, how likely one is to be sued, how easy unsafe food is to detect. Basic economics says that the minimum wage would raise unemployment in normal cases—but if there are monopsonists, it actually lowers unemployment. Furthermore, even if it raises unemployment, the harms of that could be easily counterbalanced by the wages that it raised. Figuring out whether the raised wages are worth the drop in employment is a difficult question that requires lots of empirical data and careful thought about values.

I’d say that the main core insight I gained between 7th and 10th grade about politics was that elegant political models rarely match reality. To figure out what to think about topics, one generally has to do a lot of research. It would be practically impossible to figure out the desirability of almost any policy from the armchair.

Armchair analysis can, of course, give reasons to either support or oppose a policy. A bit of basic economics is all one needs to figure out that taxes on negative externalities like pollution are generally better than direct regulations and that free trade is generally good. But when a policy has both costs and benefits, armchair speculation can’t tell you which one is greater.

Thus, foreign policy is hard to settle from the armchair. So is most economic policy and the desirability of gun control. I’ve always been amazed by the hubris of liberals who think that they can confidently predict that gun control will decrease violent crime without looking at any studies on the topic. It’s similarly ridiculous when conservatives claim to know that gun control increases crime without carefully analyzing any data on the subject.

2

To come to any confident conclusions about politics, one needs data most of the time. So how then, should we reason about big political topics—how do we decide who to vote for, for instance? There are hundreds of major differences between Trump and Biden in terms of their policy on abortion, foreign policy, immigration policy, and so on. Given that there aren’t studies looking at these things, how do we come to a rational decision about who to vote for?