Month: September 2024

I see that AH Datalytics has published the inaugural version of their Real-Time Crime Index, which provides rough estimates of crime within a couple of months of it happening. This contrasts with the FBI, which takes about six months to provide quarterly estimates and nearly a year to provide final annual figures.

What's more, the RTCI provides monthly data. For example, here's their estimate of murders over the past few years:

Does this give us any insight into why murder increased sharply in the spring of 2020? Maybe....

Up through April 2020—when COVID was already in full swing—murders remained at normal levels. They spiked in May when George Floyd was killed and then spiked further in June. But then they declined rapidly even though COVID was still surging.

So this suggests it was George Floyd, not COVID, that caused the murder spike. However, the annual summertime murder peaks have remained high for three years. It seems unlikely that the George Floyd effect could last that long, while the COVID effect probably could.

So it's still very difficult to say. If I had to guess, I'd say it was mostly COVID but with an extra little surge in 2020 when Floyd was murdered. It's just a guess, though.

This is the Belvedere Museum in Vienna. It was originally built as a palace, of course, but it never got used much. There were just too many other palaces in Vienna. Eventually, after 50 years of sporadic use, Maria Therese transferred the Imperial Picture Gallery to the Belvedere and it's been an art museum ever since.

Maybe I just read the wrong people, but I've sure seen a lot of chatter about Kamala Harris's proposal to tax unrealized capital gains. A couple of notes:

• The proposal is only for people with a net worth above $100 million. In practice it's really not a big deal.
• It's hardly unknown to tax unrealized gains. Your property taxes are based on the current value of your home even if you haven't sold it. Management fees for hedge funds are routinely based on the market value of gains, regardless of whether the underlying assets have been sold. Estate taxes are based on the fair market value of the estate's assets at the time of death. Nonqualified stock options are taxed when exercised, not when sold.
• Nevertheless, I'm not especially in favor of this. There are downsides to taxing unrealized gains, but the real reason I oppose it is that there's little point in creating a weird new tax just to hit the wealthy. At high incomes, income and wealth are almost perfectly correlated:
  .
  This means that if you want to tax the wealth of rich people, you can just establish a surtax on income above $1 million and accomplish nearly the same thing.

Why complicate things? If you think you have the votes for a wealth tax (which is what unrealized gains are), then you also have the votes for a new, higher income tax bracket for income over $1 million. So why not just do that instead?

The Wall Street Journal has this headline in today's paper:

How Immigration Remade the U.S. Labor Force

That's true enough, but it's also worth noting that it hasn't remade the workforce all that much—not recently, at least. Here's the number of foreign-born workers in the labor force:

This is all foreign born workers, both legal and illegal. It's currently exactly on its pre-pandemic trend, which suggests the recent surge in illegal border crossings has had little effect.

At the same time, it's worth noting that the share of foreign born workers has grown at an increased rate over the past year or two. This is largely because the number of native born workers has been pretty flat compared to just before the pandemic. This may be due to boomers retiring, but I'm not sure.

There's nothing new in this week's YouGov poll of the presidential race:

YouGov still has Harris two points ahead of Trump in national polling. Oddly, Harris continues to lead by four points when people are asked who they'd prefer to have as president. Perhaps the "Not Sure" vote breaks heavily for Harris?

Also, Black support for Trump suddenly jumped from 12% last week to 18% this week, while white and Hispanic support jumped a couple of points for Harris. Odd.

Paul Krugman's column today is about "left behind" places in the US. Exhibit A is West Virginia, which indeed has the lowest level of prime-age men's employment in the country:

But here's the thing: West Virginia has always been poor. Personal income in 1950 was a dismal 32% lower than the US average and central Appalachia was practically the poster boy for American poverty in Michael Harrington's famous 1962 study, The Other America. Personal income is still dismal today, but it's actually gotten a little better:

Likewise, employment in West Virginia has been lower than the US average forever. But it's also gotten steadily better. This chart is for all workers, not just prime-age men, because that's what the BLS gives us:

In 1990 the participation rate in West Virginia was 12 points less than the national average. Today it's only 7 points less.

So yes: West Virginia is poor. But it's always been poor. And recently it's gotten a bit less poor. This obviously can't explain why it's changed from a reliably Democratic state up through 2000 into a state that gave Donald Trump 69% of its vote in 2020.

So if poverty and unemployment aren't the answer, what is? Have Republicans gotten more sympathetic toward the working poor? Hardly. Have Democrats given up on them? Obamacare, the EITC, and Biden's Child Tax Credit say no.

So I guess it's just the usual racial and culture war stuff. There's certainly no "left behind" factor that can explain the change.

This has become a weirdly popular right-wing meme lately on Twitter—amplified, of course, by Elon Musk:

This interview clip is from 2019 and has nothing to do with shutting down Twitter. It took place during Trump's first impeachment inquiry after he had threatened the whistleblower who revealed his conversation with Ukranian president Volodymyr Zelenskyy—the one where he pressured Zelenskyy to dig up dirt on Joe Biden in return for military aid. Trump wanted the whistleblower treated as a traitor:

“I want to know who’s the person who gave the whistle-blower the information because that’s close to a spy,” Mr. Trump said. “You know what we used to do in the old days when we were smart with spies and treason, right? We used to handle it a little differently than we do now.”

Harris argued that this should be taken seriously and Twitter should take down Trump's account:

You have to take seriously witness intimidation. You have to take seriously an attempt to obstruct justice. You have to take seriously a threat to a witness and really to their safety and potentially their life.

....And we're talking about a private corporation, Twitter, that has terms of use, and as far as I'm concerned and I think most people would say, including members of Congress who he has threatened, that he has lost his privileges and it should be taken down.

....He does not have a right to commit a crime because he is president of the United States.... And anyone who wants to say, well, this is a matter of free speech, you are not free to threaten the life of a witness. That is a crime.

You may, of course, think that Harris was wrong to urge Twitter to shut down Trump's account. But that's all it was. She believed Trump had pretty plainly violated Twitter's terms of service by threatening official retribution against the whistleblower who had reported his misconduct. And I hardly need to point out that the whistleblower turned out to be right, do I?

I see that Rep. Jamie Raskin is trying to follow up on a weird Washington Post story from a month ago:

Democratic leaders on the House Oversight Committee released a letter Tuesday asking former president Donald Trump if he ever illegally received money from the government of Egypt, and whether money from Cairo played a role in a $10 million infusion into his 2016 run for president.

I don't know if Trump will bother responding, but if he does his answer will be "No." And that will be that.

This whole episode is kind of weird. I read the Post story about this a month ago, and wrote the following timeline about it:

September 2016: Trump meets Egyptian president Abdel Fattah el-Sisi on the sidelines of a UN meeting.

October 2016: Trump gives his campaign $10 million.

January 2017: Sisi withdraws $10 million in cash from Egypt's central bank.

"Early" 2017: CIA receives report that Sisi sought to send $10 million to Trump.

Spring 2017: Robert Mueller sets up Team 10 to investigate. They subpoena Trump bank records from 2016.

2019: Case turns cold. FBI agents ask permission to subpoena Trump banking records from 2017. They are denied.

I never bothered publishing this, and it's pretty obvious why: there's nothing there. There's a "report" that Sisi wanted to send $10 million to Trump, but no evidence that he did. And over the past month the Post has failed to move the story beyond that.

So there's maybe some slight interest here for conspiratorial types, but exactly zero in the way of concrete evidence. Until and unless that leaks out, Trump merely has to deny everything and there's nothing more to say.

POSTSCRIPT: From a purely political viewpoint, there may be some value in forcing Trump to issue a denial. Maybe. But I imagine the reason it took so long for anyone to do even that is because they were waiting for more evidence. That never came.

Rolling Stone has published a list of the 100 best television episodes ever, and there's not much point arguing over it.¹ It's a matter of taste, right?

But there's one thing that bugs me. If your reviewers are all people whose TV experience goes back only to the '80s, that's fine. I get that it's pretty tough to find a 90-something who can tell you about the great TV episodes of the '50s.

But if that's that case, then don't pretend this is an all-time list. By my count, there are only a few '70s shows on the list and a grand total of six from the '50s and '60s—and only two in the top 50.

That's not very likely, is it? Nothing—not a single episode—from Gunsmoke, Get Smart, Gilligan's Island, Mission Impossible, My Three Sons, Leave it to Beaver, Bonanza, The Addams Family, The Prisoner, Maverick, Dragnet, Bullwinkle, or Perry Mason?

This is really the Top 100 Episodes of the Past 50 Years. Why not just call it that?

¹Except for #3, that is. I liked The Leftovers, but there's no way any of its episodes, let alone the pointless, hallucinogenic one chosen, is the third best of all time. Come on.

Today is a do-over. Allow me to explain.

Nebulas in the night sky are basically just clouds of gas made luminous by starlight. They consist of hydrogen alpha gas (red), sulphur (a slightly different shade of red), and oxygen (blue). Serious astrophotographers capture this by using monochrome cameras and a set of narrowband filters that allow only the light from those three gases through. They take one batch of pictures with each of the filters and then combine them to get a color image.

You can combine the colors in various ways, called palettes. The best known is the Hubble palette, which is used by the Hubble telescope and gives its pictures their distinct look.

I can't do this because I have a color camera and those filters are expensive. However, I do have a single dual-band filter that allows only the hydrogen alpha and oxygen through. This helps the colors pop and reduces light pollution.

But today I discovered by accident that my software can mimic the palette process. You take your set of images captured with the dual-band filter and run it through the software three times, each pass extracting a different color. Then you recombine them using the palette of your choice.

I have no idea how this works. Magic? Algorithms beyond human ken? In any case, the software pulls the images apart and then puts them back together. It's not as good as doing it the real way, but it's better than nothing.

So I did this to yesterday's photo. Remember I mentioned that my image failed to produce the blue color of the bubble at the center of the Bubble Nebula? With a bunch of trial and error plus some photoshopping, I managed to bring it out. I don't think it's very good, but I imagine it could become pretty good once I learn more about this.

The top photo is the recombined image. The bottom photo is the original from yesterday. You'll notice that the top image not only captures the blue bubble, it also captures more definition throughout the rest of the nebula. This seems sort of magical since it's using the exact same data as the other picture. So where does the additional detail come from?