Month: June 2024

Retail sales were up slightly in May, but the trend is very definitely down over the past couple of years:

Retail sales per capita have declined 5% since the start of 2022. But here's more context:

So what's really going on? Are shoppers running out of money and buying less? Or simply reverting to their pre-pandemic habits? You can make a case for either one. This is why the economy is so hard to judge right now. The pandemic screwed up everything.

As part of an angry response to a New York Times explainer on getting aid to Gaza, Alec Karakatsanis says this:

For example, "humanitarian aid" pier in Gaza wasn't a cynical fraud to cover for genocide (a PR ploy by U.S. to keep sending weapons that was then itself used in one of Israel's great massacres) but a "creative solution."

I wonder how many people believe this? It's as insane as anything on the MAGA right. It's obviously true that the Biden administration is an accessory in the Gaza war, but it's just as obviously true that Biden himself is sincere in wanting to get aid into Gaza. The floating pier may have been a bit of a Hail Mary, but there was nothing cynical about it. Quite the opposite: the pier has cost upwards of half a billion dollars and has highlighted the fact that Israel is blocking land routes. That's why the pier was built.

It is plainly the case that Biden—along with nearly all of Congress—prioritizes the destruction of Hamas above food aid to Gaza. Biden has pressed Israel to do better on aid but he hasn't threatened serious consequences if they don't comply. Nevertheless, that's quite a different thing than providing "cover for genocide," which could have been carried out more easily by the simple expedient of doing nothing.

And while we're on the subject: when are we going to stop with the genocide nonsense? Based on high-end estimates, Israel has killed about 1.5% of Gaza's population. It should be possible to keep two things in your head at once about this: (a) it's horrific, (b) it's not genocide. To keep pretending otherwise is revoltingly contemptuous of true genocide.

Normally presidential candidates try to play down their prospects before a debate. That way they have a better chance of "beating expectations."

Does this work? I've always had my doubts. But we're about to get a destruction test of the theory from Joe Biden—though not from the candidate himself. From his opponent. Donald Trump has spent so much time claiming that Biden is a doddering old man that Biden will beat expectations if he just manages to stay upright for 90 minutes.

My money still says it doesn't matter. Regular people just judge on vanilla performance while political pros are in on the game. Neither cares about expectations.

What will happen, though, is a super heightened sensitivity to every facial tic and pause from Biden. Just wait and see. At the other podium, Trump will ramble mindlessly the way he always does these days, but he'll get nothing more than the usual fact check that no one pays attention to.

It's commonplace these days to say that the US doesn't know how to build things anymore. I want to push back mildly on this. Here is manufacturing:

Manufacturing production has flattened, but it hasn't declined. We make more stuff today than we did 20 years ago. Here's construction:

This is everything: houses, factories, roads, etc. Again, we build more than we did 20 years ago.

Now, this isn't the whole story. The US, for example, manufactures almost no high-tech equipment. Semiconductor production is stagnant and has lost market share regularly—though that's likely to improve over the next few years. We've had trouble ramping up defense production. And in the construction business, you can find lots horror stories of high costs and long build times—though mostly in New York City. Beyond all this, we've certainly lost ground to China:

Still and all, the US continues to manufacture a lot of stuff and build a lot of buildings. We aren't quite the flailing giant we're sometimes made out to be.

Apropos of nothing in particular, here's a list of the richest non-OPEC countries in continental Africa (no islands):

I was curious to see if the few countries that had never been European colonies did any better than the rest. It doesn't look like it. In fact, aside from oil, the main thing to look at is how far each country is from the equator:

Sub-Saharan countries are almost all desperately poor. The countries in the north and south are also poor, but not nearly as much or as consistently.

Along the Höllgasse Artists Alley in Passau, the local artists have painted a center stripe of the cobblestones in bright, primary colors. By chance, the annual painting took place the day before we were there, so the cobblestones were nice and shiny for us.

Total compensation for workers—wages plus benefits—rose in Q1:

Total comp came to $46.14 per hour. Wages and salaries clocked in at $31.72.

This is good news all around. Compensation was up 2.6% at an annualized rate, which is good for workers. But it's slowed down from 3.8% over the past year, which is good news for the Fed.

What did we learn from the COVID-19 pandemic? What should we be doing now to prepare for the next pandemic? Which measures (aside from vaccination) were most effective?

We still don't know for sure, but there's enough evidence to make educated guesses about a lot of interventions. I did this same exercise last year and not much has changed since then. Here's a brief rundown.

Intervention	Good idea?	Details
Masking	Yes, but...	Masks are required only indoors and need to be well-fitted N95 masks. Cloth masks and medical masks are only minimally effective.
Better ventilation	Yes	We should be working right now on a Manhattan Project to install better ventilation universally to prepare for a future pandemic. This is especially true in schools, workplaces, and other high-traffic areas.
Close schools	Probably not	This is a little tricky. The evidence suggests that school closures reduced the spread of COVID, but only moderately and only for the first couple of months. However, this is largely because COVID didn't affect children much, and this might not be true of the next pandemic. Also, learning loss was the same everywhere, even in places that re-opened schools quickly. Remote learning wasn't the culprit.
Remote work	Yes	Most of the evidence suggests that "essential workers," who couldn't stay home, died at 2x higher rates than others.
Far-UVC lighting	Yes	This deserves more attention. Far-UVC lighting has been well studied and turns out to be effective at killing viruses without affecting humans. We should have a crash program to install them in workplaces and other public places everywhere.
Restaurant shutdowns	Yes	Studies show that food service workers died at fairly high rates during COVID. Outdoor dining is OK. However, we might make exceptions for indoor dining at restaurants with good ventilation and Far-UVC lighting.
Ban on large gatherings	Yes	Crowded indoor gatherings seem to have frequently acted as superspreader events.
Surface cleaning	No	There were approximately zero cases of COVID that were traced back to touching an infected surface.
Handwashing	Yes	It's not clear how effective this was, but it's low cost and does no harm.
Border closings	No	There's little evidence that these helped. You can't keep viruses 100% out, and even a few cases are enough to eventually infect a whole country.
Social distancing	Yes but....	It helps, but most countries adopted a 3-foot rule that seems to have worked fine.
Contact tracing	No	It's a good idea in theory, but in practice it's all but impossible to implement well. This is especially true in a country like the US, where so many people are skeptical of government bureaucrats calling on the phone to ask personal questions.
Lockdowns	Maybe	Lockdowns were moderately effective but came at a large economic cost. The right balance is more a political question than a scientific one.

The FTC went after Adobe today for "deceiving consumers by hiding the early termination fee for its most popular subscription plan." This might be a tough one to win. Here is Adobe's online order form:

This is very similar to the page shown in the government's complaint. It could certainly provide more information, but it's nonetheless fairly clear that there's a fee if you cancel the annual plan after 14 days. Alternatively, you can pay a higher rate on a monthly basis and cancel anytime with no fee.

Now, it's true that Adobe doesn't tell you how big the fee is. The answer is that it's half the remaining amount owed, so if you cancel after one month you owe $330. That's a lot! But if you pay the annual amount up front you're also stuck even if you decide you don't want it a month later.

The FTC also complains that if you want to cancel anyway Adobe makes it a pain in the ass. I don't doubt this for a second since every single software subscription I've ever tried to cancel does this. I would dearly love a federal regulation that basically says "cancellation needs to be functionally identical to purchase and displayed right alongside it." It's the kind of nitpicky regulation I'm normally not fond of, but the cancellation merry-go-round pisses me off so much that I'd make an exception for this.

In any case, Adobe makes it pretty clear that there's a cancellation fee for the annual plan, and their cancellation process is pretty much industry standard. This is probably not an easy case to win.

When I was out at Palomar astronomizing a couple of weeks ago, I set up my camera to do some Milky Way photography. I didn't have any special equipment. Just my camera and a tripod, which limited my exposure time to about 15 seconds and also limited the number of frames I could stack.¹ Nonetheless, it turned out surprisingly well. It's a composite photo: the sky is a stack of ten frames while the foreground is a single frame lit by flash.

Next month should be the height of Milky Way season, and I have a whole different idea planned, complete with equipment that's up to the task. But that's all I'll say. After all, I might screw it up.

¹Normally you can stack as many frames as you can feed into the software. However, without equatorial tracking the top of the frame moves at a different rate than the bottom. Feed in too many frames and the software can't find a single way to register them all correctly.