Month: February 2024

Joshua Keating asks today, "what does deterrence even mean, really?" It's a good question for at least two reasons: (a) deterrence against what? and (b) what's the evidence?

The first question is critical. The most famous case of deterrence is the MAD doctrine during the Cold War. It worked! But in the modern world deterrence more commonly means the ability to stop actions short of war, such as terrorism, limited missile attacks, and so forth. The most recent example of this is a clear failure: Hamas's October 7 attack on Israel obviously wasn't deterred even though Israel has a fearsome reputation for revenge.

So even if we accept that countries think twice about starting wars they might not be able to win—like, say, Lebanon invading Israel—there's still the smaller stuff to consider. This is where evidence becomes so important.

This is pretty easy to understand. For example, Donald Trump assassinated Qasem Soleimani, the head of Iran's Quds force, and his supporters like to say this kept things quiet for a while. But did it? The evidence is pretty equivocal:

Days after the attack, Iran retaliated by firing about a dozen ballistic missiles at two U.S. bases in Iraq. While no U.S. service member was killed, more than a hundred American troops suffered brain injuries as a result of the Iranian response.

....“There have been subsequent incidents including in March when Iran-backed Iraqi militia rockets killed two Americans and a Brit. The Iranians also used their small boats to swarm U.S. Naval vessels in April — without any apparent U.S. response,” [Barbara Slavin] said.

So maybe it worked and maybe it didn't. Can you ever really know? This gets us to the big problem of empirical deterrence research. If something happens, it's easy to say that it wasn't deterred. But if nothing happens, does it mean something was deterred? That's much harder to say.

This sets up a huge bias in favor of ever more deterrence. You can always point to attacks as evidence that we need stronger deterrence, but it's exponentially harder to point to nothing—or, harder still, a little bit of something but not very much—as evidence that we already have plenty of deterrence.

But after 9/11, the US made no mystery of its purpose. It was not to deter al-Qaeda, it was to destroy al-Qaeda. Ditto for Israel after the Hamas attack. In both cases, the countries involved decided that their foe could not be deterred.

And that seems to be the case generally for guerilla wars, insurgencies, and terrorist attacks. Even the threat of massive retaliation doesn't stop them. Nation states can be deterred from flatly starting a war with a superior power, but neither they nor anyone else show much potential for being deterred from anything else.

Apparently the damages against Donald Trump in his business fraud case have been accumulating interest at a 9% annual rate for several years:

• On profits from the sale of the Old Post Office: interest is owed since May 2022.
• On profits from the sale of Ferry Point: interest is owed since June 2023.
• On everything else: interest is owed since March 2019.

This all comes to $98.6 million so far, bumping Trump's fine to $454 million. It will continue rising by about $2.7 million per month while appeals are ongoing, which means Trump is likely to owe more than $500 million if he ultimately loses his appeals and the amount of the fine stands.

TRIVIA FACT OF THE DAY: New York Attorney General Tish James, who brought the fraud suit against Trump, is exactly one day older than me.

Alex Tabarrok links today to an op-ed in the Wall Street Journal which has some good news: the T-Mobile/Sprint merger in 2020 wasn't anti-competitive! But this paragraph piqued my curiosity:

Average monthly mobile subscription fees dropped sharply. In the three years before the merger, according to government price data, mobile charges declined in real terms by about 8%. In the three years following the merger, the real price decline has been nearly 12%.

"Three years" might not sound like anything special, but in fact it's very carefully cherry picked. The merger took place in April 2020, so three years before that is April 2017. And it turns out that right in the previous month the cost of wireless services dropped a bunch. This means that if you count from April, you miss the big decline.

This is very convenient since it produces the worst possible pre-merger baseline you can squeeze out of the data. If, instead, you compare the whole period after April 2020 with the same number of months before April 2020—a much more natural comparison—you get a price decline of 21% before the merger and 16% after:

The chart above makes this clear by giving you a long-term look. Wireless phone service costs have been declining for a long time, but they've been declining less since the T-Mobile merger with Sprint.

The moral of this story is: Never trust anything on the Wall Street Journal editorial page. There's always jiggery pokerey of some kind. Always. You just have to look.

NOTE FOR DATA NERDS: I know that, technically, this ought to be a log chart. However—oh hell, forget it. Here's the log chart:

It doesn't make much difference. But it does make the slowdown in price declines a little clearer.

AND ONE MORE THING: Causality is genuinely hard to demonstrate, and counterfactuals are tough too (how much would prices have declined if the merger hadn't taken place?). So this data doesn't say a lot about whether the merger should have been approved. But it does say that the merger (a) didn't lead to lower prices and (b) most likely led to higher prices.

The judge in Donald Trump's fraud case has returned a damage assessment of $355 million. That's a lot!

This will be appealed, of course, so Trump won't have to pay anything soon. Still.

I'm not sure there's anything really new here, but today the New York Times reports on Donald Trump's view of abortion:

Former President Donald J. Trump has told advisers and allies that he likes the idea of a 16-week national abortion ban with three exceptions, in cases of rape or incest, or to save the life of the mother, according to two people with direct knowledge of Mr. Trump’s deliberations.

Mr. Trump has studiously avoided taking a clear position on restrictions to abortion since Roe v. Wade was overturned in the middle of 2022, galvanizing Democrats ahead of the midterm elections that year. He has said in private that he wants to wait until the Republican presidential primary contest is over to publicly discuss his views, because he doesn’t want to risk alienating social conservatives before he has secured the nomination, the two people said.

Donald Trump, the guy who speaks his mind and always lets you know where he stands, is suddenly shy when talk turns to abortion. How about that?

Trump could hardly be more clear that he couldn't care less about abortion. He's "very" pro-choice until he decides to become a Republican, then he switches. He says we should jail women who get abortions until, oops, he discovers that isn't the party line. He's in favor of strict abortion bans until the Supreme Court makes that possible, at which point he starts dithering about 16 weeks.

He's obviously in favor of whatever abortion policy will get him elected, full stop. There's nothing else there.

Here is Hilbert puffing his face out and sniffing the atmosphere way up on the second story. Everything in the world needs to be sniffed, over and over and over.

I was amused by this editorial choice in the LA Times today:

The point of the piece is that we shouldn't stereotype Gen Z. So what's the illustration? A bunch of teen girls stereotypically absorbed in social media. Don't the editors at the Times know that a picture is worth a thousand words?

But set that aside. The op-ed links to a Gallup report from last year about teen well-being. The chapter titles of the report are relentlessly negative, but check out the actual data:¹

I found these results interesting because Gallup asked a slew of very concrete question. As you can see, the number of teens with actively dim views of themselves and the world around them is very small. Even social media doesn't look so bad: only 13% of teens report spending more than about two hours per day on their phones.²

At the same time, about half of these teens reported feeling anxiety or stress the previous day.

This makes me wonder just how good a marker these subjective questions really are. It's well to keep in mind that phrases like "teen angst" and "teen rebellion" are practically cliches, and have been forever. So is it meaningful that teens report a lot of anxiety these days when their answers to more concrete questions are almost never negative?

And today's teens are optimistic! 80% think they have a great future. 84% think they'll achieve their goals. 85% think their life will be pretty good five years from now (7 or more on a scale of 0-10). An astonishing 38% think they will do something that changes the world. Maybe they're a little too optimistic?³

Obviously a single survey can only tell us a limited amount. And this is a snapshot in time, so we don't know how it compares to the past. Still, it's startling how big the disconnect is between soft answers (anxiety, stress) and hard answers (life is good, I always get enough to eat). Which ones should we really trust?

¹In this chart, I'm counting teens who gave an actively negative answer. In other words, it doesn't include those who reported things like "fair" or "neither agree or disagree." The point here is to find out how many teens are truly unhappy with life, not just "meh."

²The data is reported in bins (1-5, 6-10, etc), which is why it's "about" two hours.

³Not everything was rosy. 72% of teens want to get a BA or graduate degree, for example, but nearly a quarter are afraid they can't afford it. Click the link and look at the full set of questions yourself if you want a fuller picture.

Every five years the Department of Agriculture releases a Census of Agriculture. The latest edition came out on Tuesday, and for some reason I thought it would be interesting to document the decline of the small family farm. Here it is:

There's no decline at all. I've been hearing for practically my entire adult life that small farms are disappearing, but it's not true. It's medium-sized farms that are going away:

Now, it's worth pointing out that if you look at income levels, every size of farm has declined since 1997, even the smallest ones. There's only one exception: farms producing $500,000 or more, whose numbers have skyrocketed 150%.

I don't have anything special to say about this. I'm just filing it under "stuff I didn't know until today."

Good ol' Fox News:

The informant in question is Alexander Smirnov, who retailed a preposterous story to the FBI about Burisma paying bribes of $5 million each to Hunter and Joe Biden for "protection." James Comer and Chuck Grassley got wind of this and demanded a copy of the FBI interview so they could investigate whether Smirnov's accusation was being taken seriously. The whole thing took the right-wing noise machine by storm.

Today Smirnov was indicted for lying to the FBI. The indictment came from David Weiss, the special prosecutor in the Hunter Biden case. Smirnov was arrested at the Las Vegas airport after arriving from overseas.

Hard to believe, isn't it? It seemed like such a promising lead, something that could put a bow on the whole Biden Crime Family story. But no one is giving up. Biden is guilty of something, by God, and you may rest assured that Comer and Hannity won't rest until they find it.

POSTSCRIPT: Chuck Grassley is already implying that something is fishy:

"This indictment isn’t enough — the public has a right to see all the underlying evidence supporting the Biden Justice Department’s case. The Biden administration must show its work."

That's what a trial is for, Chuck.

This is a little boat landing on the outskirts of Giverny. The stream is the Epte River, which runs along the entire southern edge of town. It doesn't look like much, but it's 70 miles long and empties into the Seine a few miles from this spot.