Month: September 2024

Republicans are chattering once again about imposing work requirements as a condition for receiving Medicaid. The CBO has estimated the effect this would have:

This option would decrease federal spending...about $15 billion per year, on average, because about 2.2 million adults would lose Medicaid coverage. That reduction in enrollment represents a substantial portion of the adults who would be subject to the work requirements.

This represents a cut in Medicaid spending of less than 3%. Aside from symbolism and performative cruelty, what's the point?

According to the Washington Post, here is the partisan lean for all men in the US with names that start with "Ke":

I'm pleased to note that Kevin is the most bipartisan name of the bunch, leaning only slightly Republican.

Oddly, Kenneth is +14 Republican while Kenny is -15 Democratic. Apparently there's something Republican about insisting on using your full name.

Here's a timeless story about the way a widely cited pseudo-statistic managed to worm its way into the discourse. It all starts with a brief funding request 20 years ago that describes the outcome of a program among the Kalinga and Umili peoples in the Philippines:

2004: "From 1990-1996, a total of seven indigenous communities in Kalinga — with 1,071 households or families — were served and have benefited from this SIPAT program.... Because of this project, 81% of regional forest lands were maintained in the Province of Kalinga.  For this reason, the old growth forests remain largely intact in Kalinga and continue to contain rich biodiversity."

This is the original claim: one of the benefits of a single small program in the Philippines was maintenance of most of the forest lands in the area. This was picked up the following year in a World Resources Institute report:

2005: "The combination of watershed protection and good irrigation management raised annual incomes for over 1,000 poor families in seven indigenous communities by an average of 27 percent, all while maintaining over 80 percent of the original high-biodiversity forest cover."

This is an accurate enough paraphrase but highlights the fact that the Kalinga forest is rich in biodiversity. This apparently led a World Bank report to go a good bit further:

2008: "Traditional Indigenous Territories...coincide with areas that hold 80 percent of the planet’s biodiversity."

We've gone from Kalinga lands in the Philippines to all indigenous territory, and from 81% of the forest to 80% of all biodiversity. After that, things took off:

2008-2023: Grist says, "Indigenous land contains 80 percent of the world’s remaining biodiversity." Amnesty International says, "Although they comprise only 6.2% of the world’s population, Indigenous Peoples safeguard 80% of the planet’s biodiversity." One Earth says, "The territories of Indigenous peoples and local communities contain 80% of the world’s remaining biodiversity."

This is all documented in a Nature article that says the bullshit version of this claim is now everywhere:

2024: "Among the 348 documents that we found to include the 80% claim are 186 peer-reviewed journal articles, including some in BioScience, The Lancet Planetary Health and Philosophical Transactions of the Royal Society, and 19 news articles targeted at a specialist audience."

In a classic game of telephone, we went from 81% of regional forest being maintained by the indigenous people of Kalinga to 80% of all biodiversity in the world being maintained by indigenous communities. There's literally nothing to back it up. And just to twist the knife, the authors of the Nature article note that there's no way to measure biodiversity anyway. The 80% thing would be a pointless statistic even if it were based on something. Which it isn't.

What a classic bit of nonsense.

The whole Haitian cat-eating hoax has highlighted one of the most stunning things about the Trump era: whenever Trump makes up a story, even one that's absurd and quickly debunked, his allies immediately start up a desperate attempt to prove that he was right anyway.

In the case of Springfield, we get a picture of some guy, somewhere, carrying a dead goose. We get a stupid video of someone, somewhere, grilling something that's obviously chicken but looks vaguely like a cat. We get viral memes about Trump protecting the nation's cats and dogs. We get stories about Haitians being "dropped" in massive numbers on an unsuspecting Springfield (they came because the town wanted them to fill jobs). We get claims that a Haitian immigrant murdered an 11-year-old boy (it was a traffic accident). We get claims that the Haitians in Springfield are illegal immigrants (they aren't). Or that they were let in by Joe Biden (most have been in the US far longer). In Springfield itself we get death threats and bomb warnings against anyone with dark skin. We get claims of a huge surge in crime based on misunderstanding the FBI figures. We get claims that Haitians all have low IQs and have turned Springfield into the third world. We get J.D. Vance finally just admitting that "If I have to create stories so that the American media actually pays attention to the suffering of the American people, then that’s what I’m going to do."

None of this stuff is remotely true, but Trump's fans live in a universe where everything he says has to be true, by definition. Even if it obviously isn't, they spring into action on talk radio, on Fox News, on social media, and elsewhere to flood the discourse with bizarre nonsense meant to show that Trump was right in a "deep" sense even if he wasn't right in an actual sense. It's scary.

Here's the uninsured rate during the Trump and Biden presidencies among the non-Medicare population:

Despite a good economy, the uninsured rate rose 1.0 percentage points under Trump even if you give him credit for the pandemic decline. Under Biden, the uninsured rate has declined 1.5 points so far.

Time for something weird. Springfield, Ohio, as you know, is currently famous as the site of the Haitian cat-eating hoax being promoted by Donald Trump and J.D. Vance. This has turned out to be a fabrication, but there's broad agreement about the background story: Springfield was a dying town until 2014, when local leaders began a project to attract new employers and, eventually, new workers. Many of the workers were Haitians living in Florida and elsewhere who heard the town's appeal and moved to Springfield in hopes of getting better jobs. In the end, something like 15,000 Haitian immigrants came to Springfield.

But there's something odd going on. Here's the population of Springfield (the city itself, not the MSA):

Springfield's population has been flat since 2014, actually declining slightly. Here is employment:

This confirms the overall population story. Employment has been flat since 2014. Of course, this doesn't tell us anything about the composition of the population. Here's that:

Through 2022, at least, the Census Bureau recorded virtually no growth in the number of residents born outside the US.

This gibes with decennial census data. Between 2010 and 2020 it records a drop of 5,000 white residents, no major change in Black or Hispanic residents, and an increase of 2,500 in multiracial and "other." That puts a hard cap of 2,500 on the number of new immigrants through 2020, and the real number is probably a good bit lower.

Housing units? Down slightly. GDP? Down slightly. School enrollment? Flat. Median income growth? A bit less than the national average. The number of people who don't speak English at home? Less than 3,000.

What the hell is going on? This data comes from the Census Bureau, the BEA, and the BLS, so it's not just a matter of a single agency screwing up. They're all producing data that's in broad agreement: there's been no economic renaissance in Springfield, no population recovery, and nowhere near 15,000 new immigrants of any nationality.

Obviously there are Haitian immigrants in Springfield. But how many? I'd be surprised if it was much over 2,000, with possibly some more living outside city limits. Either that or there's something hella weird going on with our statistical agencies.

UPDATE: Using data through 2023, the Haitian population of Springfield might be closer to 4,000. Details here.

Kamala Harris sat down yesterday for an interview with Brian Taff of Philadelphia’s Action News. He asked her this:

When you talk about bringing down prices and making life more affordable for people, what are one or two specific things you have in mind for that?

Harris is getting a lot of flak for not really answering the question. She filibustered for 3½ minutes and repeated her talking points about building more housing, providing a $50,000 tax deduction for starting a small business, and giving young families $25,000 in down payment assistance.

This is actually a perfectly adequate answer about making things more affordable, but says nothing about bringing down prices. Why?

For better or worse, it's because Harris is tolerably honest. There is virtually nothing a president can do to lower prices, and it's deceitful to pretend otherwise even if it's politically expedient. So two cheers for Harris not serving up a big dollop of BS about how she's going to lower prices for everyone.

POSTSCRIPT: Go ahead and watch the whole interview. You know the common trope that local reporter ask better questions than the bigfoot national reporters? Nah. They both suck.

The Wall Street Journal has a story today about consumers becoming delinquent in paying their bills. The story isn't wrong: both credit card and auto loan payments have transitioned into delinquency (30 days late) at increasing rates over the past couple of years.

The data comes from a quarterly Fed report, and the problem is that it's very hard to conclude anything from it these days. If everything were headed in the same direction, it would be easy. But it's not. For example:

If you look at severely delinquent loans, there's nothing much going on with autos and only a moderate problem with credit cards. But there's more:

The actual overdue balance on all consumer loans has been steadily down, with only the tiniest of upticks in the past year. And there's this:

Loans in collection have trended down for years and show no problems at all.

The Fed's credit report has been like this for a while: depending on what you look at things can seem good or bad. You might think of the credit card data as a canary in the coal mine—it starts to go bad first and signals trouble ahead—or it might be that it's just the most volatile signal and should be discounted unless it's been bad for a while.

I don't know. Overall, credit trouble does seem to be increasing over the past year, but only slightly. It could be a sign of trouble ahead; a reversion to the mean after pandemic rescue funds dried up; or just a slight wobble ahead of a soft landing. Take your pick.

Ezra Klein put up an interview yesterday with our real border czar, Secretary of Homeland Security Alejandro Mayorkas. It was interesting. Mayorkas talked about several causes of increased illegal immigration, the prime one being the evolution of highly professionalized smuggling organizations. He also mentioned wars and crime surges in certain Central American countries, which have motivated more people to leave. And there's the huge and growing backlog of asylum cases, which has made asylum requests a free ticket to at least 5-10 years in the US while waiting for a judge to decide your case.

But one thing that neither he nor Ezra mentions is the simplest one of all: jobs. When jobs are plentiful, we get more illegal immigrants. When jobs are scarce, we get fewer.

I've mentioned this before, but here's yet another way to look at it. If there are more job openings than actual hires, it means US businesses are urgently in need of workers. So let's look at the difference between openings and hires over the past decade:

Illegal immigration follows the job market almost perfectly. But not quite: there are still spikes that happen for other reasons. The 2019 spike was likely due to the start of professional smuggling organizations and ended quickly because of Donald Trump's brutal family separation policy. The 2023 spike, according to Mayorkas, was due to a temporary gap in Mexican border enforcement. And the 2020 dip in jobs was obviously because of the pandemic.

So there are policy changes and other outside events that make a difference, but usually only for short periods. Generally speaking, illegal immigration is all about jobs and not much else.

I'll confess up front that I have no idea if this is legit:

I will say, however, that I'm wide open to believing it. Based on previous versions of ChatGPT plus what I've heard about o1, an IQ of 97 sounds very plausible. As with IQ scores on all standardized tests, this is an average of various subtests. On some of them it scored higher than 97 and on some it scored lower.

Speaking of, I watched Oprah's AI special last night and mostly laughed at it. For some reason it reminded me of how strongly we resist the notion that modern LLM models say more about us than about AI.

Sam Altman said on the show that ChatGPT "just" examines a long string of words and then predicts the next one. "Like word suggestions on your phone," Oprah chirped. Yes, Altman replied, except more sophisticated.

This makes it sound like LLMs are just a parlor trick. And I suppose they are. But they also sound remarkably human. That's because taking in context and mechanically creating a response is also how human brains work. There's very little real thinking or understanding going on most of the time. We just aren't nearly as sophisticated or creative as we like to believe.

And that's why AI is certain to get really good really fast—compared to humans anyway. Not because it's all that smart, but because we aren't.