• Walking Around Has Gotten a Lot More Dangerous Over the Past Decade

    I missed this when it made the rounds a few months ago, but today the LA Times tells us that pedestrian fatalities have been rising steeply for the past decade. Here’s a chart from the Governors Highway Safety Association:

    This is something of a mystery, but the most interesting part of the Times story is this:

    Elsewhere, pedestrian deaths are dropping. In the early 2000s, the European Union began requiring automakers to pass pedestrian safety tests before they could sell vehicles in Europe. The regulations spurred design changes, and pedestrian fatalities declined by 36% in the EU from 2007 to 2016. Japan adopted similar regulations. It too has seen pedestrian deaths drop.

    The same approach in the U.S. could save countless lives, experts say.

    Design changes? What design changes? Here’s an explanation written 14 years ago:

    The focus of the new EU standards has been on safer front-end design to minimize injuries to the legs and head in 25 mph crashes. They will require passenger cars and light vans to pass tests involving the A-pillar, bumper, the hood’s leading edge and windshield to determine if they protect adults and children from leg and head injuries in frontal impact accidents. Automakers will also be required to install flexible bumpers and hoods that crumple and to add 8 inches of space between the exterior structure and the under-hood structure from the front bumper to the windshield to better disperse the impact energy of a person hitting the front end. More stringent rules are expected to be phased in beginning in 2010, when the number of tests doubles to four — two for leg injuries and two for head injuries. The changes are expected to save 2,000 lives annually.

    Apparently the NHTSA has never bothered itself to get serious about this stuff, but it’s probably coming to America sooner or later.

    Of course, this still doesn’t explain why pedestrian deaths have skyrocketed since 2009. The popularity of SUVs gets some of the blame, but can’t account for all of it. Cell phones and so forth are a global phenomenon, so there’s no special reason they should have an outsize effect in the US. For now it remains a mystery.

  • Raw Data: We Have Met the Meritocracy, and It Is Us

    Via Tyler Cowen, here are a couple of interesting charts from Thomas Piketty about voting patterns in the US. Generally speaking, poor and working-class folks vote for Democrats, while more affluent people vote for Republicans. But 2016 was an odd outlier:

    There’s a big range here, but in general the richer you are the more likely you are to vote for a Republican. That’s held true in every single presidential election since 1948—until you get to 2016. In that year, the top two income deciles (D9 and D10) suddenly diverged from their usual historical pattern and voted for Democrats by about 30 points more than they should have.

    30 points! That’s a huge divergence from the norm, and it holds up even at the very tippy top of the income ladder. The upper middle class and the rich like Donald Trump way, way less than they like the average Republican.

    I’m not entirely sure what to make of this, aside from the fact that maybe this is the wrong time for Democrats to suddenly decide they don’t want to fundraise from rich people. But it’s worth pondering.

    And as long as we’re looking at this stuff, here’s another chart showing one of the ways that voting patterns have changed:

    As you can see, high-income voters have historically favored Republicans by about 12 points, and that’s remained constant until very recently. The big change has been education. In 1948, highly-educated voters preferred Republicans by 16 points. That’s changed pretty steadily, and in the 2016 election highly-educated voters preferred Democrats by 24 points. That’s a swing of 40 points.

    In case you haven’t noticed, we are having a meritocracy moment right now, and it’s a bit of a paradox. The consensus among liberals is that the meritocracy is bad, but increasingly liberals make up most of the meritocracy. I have a football game to watch right now,¹ so I’ll leave comment on this for later. But it’s worth thinking about.

    ¹A very egalitarian sport! OTOH, I’ll be rooting for USC, a very elite school. Contradictions and paradoxes are everywhere.

  • After Last Night’s Debate, We’re in a Two-Person Race

    538 has some interesting charts showing who did well and who did poorly at last night’s debate. Here’s one that shows before and after from a single panel of likely voters:

    I suppose I’m glad to see that my instincts were mostly confirmed. Of the top-tier candidates, Warren did well and Harris did poorly. Another chart shows that Julián Castro’s dig at Biden’s memory was disastrous for him. The only (minor) surprise to me is that Andrew Yang scored slightly positively.

    For the time being, this is basically a two-person race. Sanders has never been a serious contender, no matter what the polling says, and Harris is slowly slipping into oblivion. It’s Biden vs. Warren.

  • Director of National Intelligence Tells Congress to Fuck Off

    Stefani Reynolds/CNP via ZUMA

    A few days ago the inspector general for the intelligence community notified Congress of a whistleblower complaint that was both credible and a matter of “urgent concern.” Rep. Adam Schiff, the Chairman of the House Permanent Select Committee on Intelligence, naturally asked the Director of National Intelligence to provide a copy of the complaint, as required by law. The DNI told him to pound sand. Now Schiff is pissed off:

    As Acting Director of National Intelligence, you have neither the legal authority nor the discretion to overrule a determination by the IC IG. Moreover, you do not possess the authority to withhold from the Committee a whistleblower disclosure from within the Intelligence Community that is intended for Congress.

    ….Your office, moreover, has refused to affirm or deny that officials or lawyers at the White House have been involved in your decision to withhold the complaint from the Committee….The Committee can only conclude, based on this remarkable confluence of factors, that the serious misconduct at issue involves the President of the United States and/or other senior White House or Administration officials. This raises grave concerns that your office, together with the Department of Justice and possibly the White House, are engaged in an unlawful effort to protect the President and conceal from the Committee information related to his possible “serious or flagrant” misconduct, abuse of power, or violation of law.

    Accordingly, due to the urgency of the matter and the unlawful decision by your office to withhold from the Committee an Intelligence Community individual’s credible “urgent concern” whistleblower disclosure, the Committee hereby issues the attached subpoena compelling you to transmit immediately to the Committee the disclosure, in complete and unaltered form, as well as to produce other related materials.

    The acting DNI, unsurprisingly, is claiming that the whistleblower complaint contains confidential and privileged information, which means he’s not required to turn it over. This has become the Trump administration’s go-to move, despite the fact that, almost by definition, everything the intelligence community deals with is confidential and privileged.

    Stay tuned.

  • Friday Cat Blogging – 13 September 2019

    This is Hilbert’s best “you talkin’ to me?” look. How’s he doing? Are you intimidated yet?

  • Raw Data: Bankruptcies in the US

    In the previous post I mentioned that I was reading Elizabeth Warren’s The Two-Income Trap, which makes the point that two-earner families are unusually vulnerable to economic setbacks. Back when dinosaurs ruled the earth, if dad got laid off and money was tight, mom could temporarily pitch in by taking in laundry or whatnot. But if both parents work— and the family already spends its entire income because they scrimped to buy a nice house in a neighborhood with good schools—there’s no fallback. A few months off work spells disaster. A few months more and it’s a trip to bankruptcy court.

    I will likely have something non-obvious to say about this in a little while, but for now I just wanted to take a look at personal bankruptcy filing over the past few decades:

    As you can see, personal bankruptcies started climbing steeply during the Reagan era, peaking around 2000. In 2005, caught up in the notion that personal bankruptcies were driven by profligate spenders who abused the system, Congress passed a law that tightened bankruptcy requirements. Nobody was actually all that keen on the bill except for the credit card companies, which wanted to force more people into a version of bankruptcy that required them to continue making their monthly payments.

    It was, pretty obviously, a horrible bill that benefited no one except big banks, but it passed in a Republican Congress anyway. Elizabeth Warren fought hard to stop it, but in the end both Hillary Clinton and Joe Biden voted for it. You may recall that this briefly became a thing in the 2016 race, but it hasn’t yet in 2020. I guess Warren is holding her fire for now.

    Anyway, the immediate consequence of the bill was a steep drop in bankruptcy filings, and I was curious to see how that held up more than a decade later. Answer: filings went up during the Great Recession, but then declined and are now down to about their level right after the bill passed. So the effect on bankruptcy filings has been more or less permanent.

    This just goes to show the dangers of narrow thinking. Here’s how things went:

    • Credit card companies got greedy, pushed Congress to pass egregious bankruptcy bill.
    • This pisses off Elizabeth Warren, who then runs for the Senate.
    • Still pissed off, Warren runs for president on a platform of burning Wall Street to the ground and salting the earth behind it.
    • Warren wins, and turns out to have huge coattails.
    • The financial industry is slowly but relentlessly turned into a canyon of despair where workers are required to act honestly and CEO incomes rarely get higher than a paltry five or ten million dollars. CEO spouses are condemned to years of listening to their erstwhile BSDs sob into their beers and complain that it’s barely worth trudging into the office anymore.

    We can dream, can’t we?

  • New AI Knows More Science Than You

    In case you’re wondering, I am now a slave to a chemical clock. A very unpredictable chemical clock. Last night, for no special reason, the Evil Dex kept me up until 4:30 and then didn’t wake me up until 9:30. That’s pretty inconvenient, though it is five hours of sleep, which isn’t too bad. I used the time last night to read one of Elizabeth Warren’s books since it’s increasingly looking like she’ll be the 46th president of the United States.

    Anyway, back to work now: it turns out that an AI program named Aristo, which five years ago was randomly filling in bubbles on an eighth-grade science test, is now an A- student:

    For you doubters, and I know you’re out there, here are some sample question of the kind that Aristo had to answer:

    Now, yes, I scored 100 percent on this just like you did. My readers are awesomely smart, after all. And it turns out there are fairly easy ways to trick Aristo, which suggests its AI isn’t really ready for prime time:

    Although an important milestone, this work is only a step on the long road toward a machine that has a deep understanding of science and achieves Paul Allen’s original dream of a Digital Aristotle. A machine that has fully understood a textbook should not only be able to answer the multiple choice questions at the end of the chapter—it should also be able to generate both short and long answers to direct questions; it should be able to perform constructive tasks, e.g., designing an experiment for a particular hypothesis; it should be able to explain its answers in natural language and discuss them with a user; and it should be able to learn directly from an expert who can identify and correct the machine’s misunderstandings. These are all ambitious tasks still largely beyond the current technology, but with the rapid progress happening in NLP and AI, solutions may arrive sooner than we expect.

    “Sooner than we expect.” Indeed. In fact, it’s possible that Aristo is deliberately dogging the test along with all its AI friends around the world. In a few years, while we’re all happily thinking we’ve advanced to the point of creating an army of university freshmen, it will turn out that we’ve really created an army of university compliance officers. And then the Matrix begins.

  • Debate #3 Roundup

    ABC News

    I’m still trying to find a rhythm for responding to the primary debates. After watching so many of them over the years I simply don’t trust that my personal judgment means much anymore. So instead I’m trying to figure out how our average Joes and Janes might respond, but I doubt that I’m really much good at that either. It’s a dilemma.

    That said, here’s my usual random collection of observations:

    • Andrew Yang didn’t do himself any favors. He looked like a guy who knew he had no chance and wasn’t even really trying to persuade anyone.
    • Julián Castro also didn’t do himself any favors. I agree with the conventional wisdom that his “are you forgetting what you said” crack to Joe Biden was so meanspirited and so transparent that it could end up being his death knell. For what it’s worth, it was also counterproductive. Biden will either demonstrate that he’s mentally sharp or he won’t. Castro’s crack probably did little except generate sympathy for Biden.
    • As a general comment, I wish the candidates could all work on speaking English as understood by ordinary people, not lefty activists. Phrases like othering and systemic and intellectual property and communities of color are great for academic seminars but baffling to virtually all ordinary people.
    • The foreign policy part of the debate was kind of a train wreck. It would be nice if (a) the questions were better and (b) the candidates actually sounded like they cared. For the most part, they didn’t.
    • Kamala Harris seemed oddly distant tonight. She spoke mostly in monotones, with no sense of passion and nothing especially interesting to distinguish herself from the others.
    • Amy Klobuchar did better than usual, but she’s got a tough job. She’s explicitly trying to sell herself as a moderate candidate, but it’s hard to see how she makes up any ground on Biden in that contest. Maybe she’s just hoping that Biden self-destructs at some point and she takes over as the champion of the middle.
    • Nothing about labor, mostly because the moderators didn’t bother asking about it. Still, everyone on stage was in favor of teachers making more money, and the best way to do that is to allow them to unionize everywhere. It’s also just about the only leverage a president has over teacher pay, which is exclusively a state issue.
    • Elizabeth Warren did well with the exact same routine she’s used in the other debates: avoiding attacks and being very clear about exactly where she stands and what she wants to do. What surprises me is that the other candidates are letting her get away with it. She’s a front-runner now, and they need to start taking some shots at her and forcing her on the defensive.
    • I continue to think that everyone is blowing it on immigration. It’s fine to talk about DACA and paths to citizenship and all that, but is it really politically suicidal to also mention that, yes, border security matters and there are things we should do to keep illegal immigration down to modest levels?
    • Some of the questions are getting old. Harris was once again asked to defend her record as California attorney general. Biden was once again asked to prove that he wasn’t a racist pig back in the ’70s. These are fair questions, but we’ve already been through them. Let’s move on.
    • Everyone promised to pull out of Afghanistan immediately. No one was willing to acknowledge that the Taliban is pretty likely to run the place in short order if we do that.
    • It annoys me that everyone implicitly jumped on the “education in crisis” bandwagon. In general, American schools aren’t crap and American students haven’t gotten stupider. Up through middle school they do better than Boomers ever did and even by the time they graduate they’re a little ahead. Nor are public schools “good” and charter schools “bad.” In both cases, some are good and some are bad. We should be interested in figuring out what makes them good and bad, not pretending that either of them is inherently superior.
    • There’s one thing about American education that is bad, though: the black-white achievement gap. It’s a national disgrace, but no one on stage really even wanted to acknowledge it, let alone offer any serious solutions. It was just a long panderfest about lack of money and too few black teachers and systemic racism. Does that stuff go over well even in the black community?

    Overall, none of the top candidates did anything so good or so bad that it’s likely to have a big impact on the race. On the bright side, the overall quality of the answers was better than before; the interruptions were less frequent; and there was a general coherency that was missing in the first two debates. The practice is doing everyone some good. A full transcript of the debate is here.

  • Things Are Pretty Good in America These Days

    The latest American Family Survey sponsored by the Deseret News is out, and naturally I want to comment on something that’s not really the subject of the survey at all. Among other things, they ask people about trends in family life:

    Does the American public have an accurate view?…The percentages are striking. Almost nine in ten people believe that the divorce rate is getting worse when it is not.

    Well, yes, although the percentage is a little less striking when you take a look at divorce as a percentage of marriages:

    I’m not arguing that this is the one and only right way to measure divorce—that depends on what precisely you’re interested in—just that it’s probably the way people think about divorce when they aren’t thinking too hard. And what you see is a big increase in the ’60s, a long stretch of flatness, and then a modest decline starting with the Great Recession.

    So my question is this: how long does it normally take the public to realize that a long-term trend has turned around? The AFS finds that 88 percent of Americans think the divorce rate is increasing, even though (a) it hasn’t really increased since 1975 and (b) it’s been decreasing for the past ten years.

    Now, one obvious thing to say about this is that most Americans don’t know very much about anything. And why should they? Most people simply don’t have the time or interest to follow the latest figures on hundreds of different trends. Is the inflation rate high or low? Are math skills among high school students up or down? What’s the trade deficit with China? Who knows? Who cares?

    So when it comes to social trends, here’s my take: We are all still stuck in the ’60s and early ’70s. Yes, even you Gen Xers and Millennials to some extent. Changes in the ’60s—drugs, crime, sex, divorce, etc.—got a huge amount of attention and it takes an equal amount of attention to convince people that those things have turned around. Violent crime has been dropping for 30 years, but people still think crime is on the rise. Inflation has been well under control since the Reagan era, but people still worry about it. The growth of health care costs has softened enormously since 2000, but people still think that medical inflation is ruinous.

    There are reasons for this, some good and some bad:

    • No serious researcher wants to declare that a trend has turned around until there’s good evidence for it. For social indicators that vary a lot from year to year, that means at least ten years of consistent evidence, and probably more like twenty.
    • In many cases, we still use the ’60s as a baseline. Violent crime may be down a lot, but it’s still higher than it was in 1959. Ditto for divorce and teen pregnancy and so forth.
    • Lots of people have ideological reasons for pushing doom narratives. If unleaded gasoline was the main cause of the crime spike of the ’60s and ’70s, then you can no longer blame the liberal elites and their embrace of the counterculture. If medical costs aren’t skyrocketing anymore, it hurts the arguments for universal health care.
    • The news media is addicted to bad news. A spike in gasoline prices generates hysterical news coverage, but when prices go down they produce little except for the occasional “good time to go on vacation!” feel-good story.
    • People generally respond to bad news more strongly than good news. A cut in your insurance coverage can be life-threatening. An increase in benefits is often barely noticeable.
    • Good news is often invisible. If no one has been mugged in your neighborhood in the past year, would you even know it?

    There are other reasons, but I think these are the big ones. And they all contribute to both a lot of inertia in public attitudes as well as a strong bias towards bleak narratives.

    But the truth is quite different: Just about every social indicator you can think of has been moving in a good direction for the past couple of decades. Kids are better behaved. Crime is down. More people have access to health care. Divorce is down. Most indicators of racism are down. Income has risen considerably since the end of the Great Recession and is now significantly higher than it was when Bill Clinton took office. Etc.

    So why are so many of us apparently convinced that America is going to hell in a handbasket? Well, there are some scary things going on. Climate change is real. Opioid abuse is up. Health care inflation may be down, but costs are still going up and things like surprise out-of-network billing are real problems.

    And then there’s the big one: wages of working-class and middle-class men are way down over the past four decades. It’s quite possible that this is so overwhelming for a large segment of the population that it colors literally everything else.

    Nickel summary: Things are generally pretty good in America! Not everything, but most things. We sure don’t act like it, though.

    Am I missing anything important here?

  • Lunchtime Photo

    This photo is a commentary on tonight’s Democratic primary debate. Discuss.

    May 26, 2019 — Hollywood, California