Thursday assorted links


Soltas' analyais is very good, but I'm not convinced. I don't think anti-trade sediment must be correlated with the likelihood that one is personally impacted by free trade in the manner that he supposes.

I live in Michigan and I don't think Soltas' analysis is very good at all. Looking at the county-by-county results, Trump did best in the areas that have suffered from de-industrialization and globalization over the past few decades. I don't think he's measuring the right data. He seems to have done worst in counties dominated by a university or are headquarters for multi-national corporations.

Thanks for that. I understand what he was trying to measure, but I did not have any way (at least in the 3 minutes it took to read it) to tell if he was indeed measuring the right thing.

"Looking at the county-by-county results, Trump did best in the areas that have suffered from de-industrialization and globalization over the past few decades." So that is explicitly what I checked in my analysis, and it's false. Disagree? Show me the data.

Looking at the data, I think the problem is that you're confusing two ideas: the counties that had the largest drop in manufacturing and "areas where manufacturing's decline has been least important." That's not the same thing. Look at Ionia and Jackson, two counties in which Trump did relatively poorly. Both had large declines in manufacturing jobs, but both are also how to large state prisons. A quick look shows me that most of the counties with large manufacturing declines that didn't for Trump have more diversified economies. The ones that I know lost both jobs and have few other options went for Trump.

To be clear, here is precisely how I measure manufacturing. I use take the change in the number of manufacturing jobs from 1990 to 2007 in a county and divide that by the size of the working-age population in that county. (It's actually David Dorn's calculation, not mine.) So this, I think, captures exactly what you are claiming about diversified industrial bases. Those counties with diversified economies that were most resilient to the economic decline were also the strongest for Trump.

That's how are you showing they are diversified? A decline in manufacturing jobs over a period like that can mean an increase in diversification.
Take Kent County, which would seem to prove your theory. They lost a lot of manufacturing jobs between 1990 and 2007 according to your data. But the area's unemployment rate is at 3.5 percent. Why? Because it has added a lot of meds and eds jobs since then. Neighboring Ottawa county did just as bad in manufacturing loses, has an even lower unemployment today and cast fewer votes for Trump.

By contrast, Ontonagon went big for Trump. Nothing has really replaced the 182 jobs lost when Smurfit-Stone closed down in 2009 and it has had the second highest wage drop of any county in the state since then. The population has declined by about 10 percent. The high school had to drop is auto shop, meaning the potential for future high paying jobs has diminished. These folks actually benefited from trade, but now they feel screwed.

There are even more ways you're wrong because you don't have any context, but I think this is enough for now.

About 2/3 of the manufacturing jobs left in MI are in export industries and they would be lost without any guarentee that any jobs twould come back to MI if we stopped trade since wages are less in other states. C

I'd like to see a paper on why young people start so many sentences with "so." Has it replaced "like" or "um?" That might make sense in oral conversations, but I see it in written text all the time.

I'm not convinced either! It's just one blog post. And I agree that manufacturing's decline might not be correlated with protectionism. But, two points. 1. That is the story people have been selling. 2. Can't think of any other local measures of potential support for protectionism.

#3 Given the deep, brutal crisis of replication in psychology and social science, I think such articles should be mostly ignored unless preregistered etc.

I thought you read Andrew Gelman's blog.

Could any gender stereotype be stronger than the "all husbands are knuckle-dragging idiots" stereotype that invaded advertising about 15 years ago? It's starting to seem like every single product from every company includes at least a partial nod to this foolishness. An easy way to spot it is by switching genders: would the commercial or print ad be viable if you reversed the genders of the participants?

Probably it's effective because women make quite a lot of independent decisions about many areas of household spending, and it makes them feel good to be portrayed as the sensible ones who talk sense into their husbands/boyfriends.

Yeah, were gender roles to be switched, there would be a massive PR disaster awaiting for the company, and the ad campaign would be rapidly pulled. But, it doesn't bother me. We can take it. And if women feel empowered to make all those tiresome decisions related to basic household goods (especially cleaning supplies, design and furnishings, food, children's clothes and toys...), then all the better, 'cause then all that stuff will just magically appear without the tiniest of effort (well, gotta chip in for the expense and some help with pickup, delivery and unpacking). Also, having let them make such a very large number of those essentially inconsequential decisions (which matter very much to them, but less to us), you are better positioned to insist that it is your turn to call the shots for things that actually matter to you (say, which cable package, a nifty feature for the new car, or perhaps a necessary upgrade for the man cave). Win-win, I think.

"Yeah, were gender roles to be switched, there would be a massive PR disaster awaiting for the company, and the ad campaign would be rapidly pulled. But, it doesn’t bother me. We can take it."

Translated: Sexism against men exists in popular culture, but it's okay because men are more capable of dealing with sexism than women.

Nathan, you are supposed to hide your sexism/racism of low expectations better.

Meh. I'm mostly peddling in stereotypes in this instance. It was not an entirely sincere argument. I don't fraternize with women who hold disdainful attitudes towards men, nor will I ever. Kind of like it stuns me that any woman would want to spend time around any man who holds disdainful attitudes towards women.

Ideally, both men and women would try to hold generally respectful attitudes towards the other gender, even though this is not necessarily warranted for all individuals of each gender.

Keep digging Nathan, Thomas has it in one and you agree with him: Sexism against men exists in popular culture, but it’s okay because men are more capable of dealing with sexism than women.

Correct. And then?

No, I didn't remotely suggest it's OK. In fact, I suggest completely writing off anyone who does such things, immediately and permanently excluding them from your social network - moreover, I think this should be a two way street. However, discussing legitimate issues of concern such as sexual assault or negative attitudes towards women in certain professions should be encouraged, so long as it does not promote the idea that men are GENERALLY like that.

For example, Clinton receives an awful lot of utterly disgusting commentary that would never in our current society be lobbed at a man - like, who says of a male presidential candidate "he's ugly, therefore unelectable" or makes fun of their pant suits? Men who do that sort of stuff should be ostracized, perhaps until they come around, just as much as women who peddle in jokes about how dumb men are (on-stage comedians generally excepted, since they usually do so in a way that draws attention to the negative stereotype, rather more so than in fact perpetuating them).

So, one of the main reasons to avoid stereotyping women is to avoid contaminating young girls with bad social messages. How does that not apply here? Is it OK for boys to grow up hearing the message that men are hopelessly stupid, insensitive, looser husbands? If they hear that their whole lives, is it OK if that's how they turn out?

For the understanding impaired: I am specifically NOT saying sexism against women or girls (or even those of more fluid gender) is OK. Clearly it is not. What I'm saying is that stereotyping ANYONE is wrong.

Wow. #4 is very much a 2MA, Second Machine Age, thing. "Give junior employees better technology, he said, and they offered just as much value as a managing director would have in days gone by."

I have a theory that it was never about the employees per se, but about the incentive structure. That theory is, it used to be that lazy managing directors were a key part of the recruiting and motivational pitch. "If you work very, very hard all night, you can become just as lazy and rich as your MD" was always untrue of course, since politics determined who became the MD (therefore perpetrating the dominance of "lazy" MDs), but junior people believed that pitch and flocked to the colors. The times have changed. The internet exposed the changing bonus pool structure since 2008, just how much toiling for nothing was going on at the junior level, and that made the I-banks unattractive for milennials. At the same time, the Valley has provided a plausible alternative for hard working types. In consequence, you don't get any recruiting traction from those MDs any more nowadays.

That is interesting, and sounds reasonable as a factor. More generally though.

"Overall, managers and executives were the most likely to think their jobs will be eliminated in 50 years’ time, but that’s probably because they think we’ll all have finally realized they’re just sitting in that glass office playing Solitaire all day."

3. Humans are biological organisms, you cannot abolish biology

But you can certainly drink the kool-aid.

For example, 100,000 years ago on the African plains, the best welders were clearly men. So clearly superior that the genetic ability for women to be welders was weeded out through evolution.

And let's not be picky about the reality that welding was not a human skill at the time - just so stories work so much better when ignoring reality.


Google "The unbearable accuracy of stereotypes"

A lot of stereotypes are driven by outliers. Gang violence and African Americans, for example, whereas 99% of African Americans are not in gangs or commit violent crimes. I think almost all stereotypes have a grain of truth, but the tendency to generalize/stereotype can be troublesome.

Harding, I was thinking a similar thing but my estimate was closer to 90-95%. I couldn't find the exact numbers I was looking for easily, though; it looks like something like 1% of African Americans are currently in prison for a violent crime, and on average 0.5% commit a violent crime each year, but given recidivism rates that's hard to turn into the number that have ever committed a violent crime. Something like one in five have prison records by the time they're 30, but those are mostly for drug crimes and it's hard to back out the number who have committed at least one violent crime.

Has someone run the numbers to answer this specific question?

Might want to do the numbers for African American males, I think that is largely where the stereotype lies

Whatever the number is, we all know it is too high. I wish some of the more outrageously racist individuals would not hijack the conversation every time people try to discuss the matter, as it makes it nigh on impossible to discuss practical solutions without someone coming along to peddle in various damaging generalizations - the response is predictable, in that people who stray more innocently into such conversations are similarly tarred/generalized for the argumentative strategies of the racists, making it difficult to even broach the topic. It is not a good situation.

I advocate for drug legalization in order to end turf wars, in response to this observation of a larger percentage of black males facing criminal sanction for their behaviours (which I assume reflects an underlying reality where they are also committing more violent crimes, even though racial profiling probably leads to the statistics over-stating the underlying difference in current behavioural tendencies).

The most significant story of the year is AlphaGo, and you guys are discussing free trade? Ha. As Krugman has pointed out, and others, even TC, trade is only at best about 15% of GDP's composition, which works out to about 1% or so on GDP's rate (details on Krugman's wonkish post). But the fate of your kid's future is in AlphaGo's progress.

Meanwhile, I'm winning vs my PC, set to low Expert. Here is the position, White to move, I'm winning I think [I am but only due to PC mistakes! Aggh!]

Ray vs PC White to move, I will play 22. Qa4 and white is winning...but Oh! I missed in my haste the best moves...I sent an hour thinking on 15. Ba6?!, and did not play the obvious 15. a6! Oh my, and the PC played 17...Qxd2? allowing me back in the game, but I missed the obvious rook and knight checkmate, as well as the queen pin on b7; I played quickly without considering the best moves, oh my! I'm a patzer, I play too slow or too fast...

1. e4 Nf6 2. e5 Nd5 3. d4 d6 4. exd6 Qxd6 5. Nf3 Nc6 6. c3 Bg4 7. Be2 Nf4 8.
Bxf4 Qxf4 9. O-O e6 10. Qb3 O-O-O 11. Nbd2 Kb8 12. a4 Rg8 13. a5 Rd6 14. g3 Qh6 15. Ba6 Nd8 16. Bxb7 Nxb7 17. Ne5 Qxd2 18. a6 Rb6 19. Nd7+ (19. Nc6+ Kc8 20. axb7+ Rxb7 21. Nxa7+) 19... Ka8 20. Nxb6+ axb6 21. axb7+ Kxb7

According to this chart US trade import/export is almost 30% GDP. This has greatly risen since US producers and labor lost influence in the 1960's70's and the international financiers started to control the policies in Washington

"This has greatly risen since US producers and labor lost influence in the 1960’s70’s and the international financiers "
Yup, it's all a policy story. Nothing else has changed in the larger world in the last forty years.

"The most significant story of the year is AlphaGo"

Uck. I'm inclined to agree with this. Everybody wants to believe that having a Democrat with bad hair win the Republican nomination is vastly important, but it isn't.

It also isn't important that computers are now superhuman in Go.

What's important is that major surprises in AI are apparently possible. Nobody would have called this two years ago. It may be a worrisome development.

In stories of the decade, it's up there with the rise of SpaceX.

Right. The thing here isn't that computers can do Go; the thing here is that human thought is simpler than we thought it was. Which means that other things that looked hard to become superhuman at probably aren't as hard as they seemed a year or two ago.

Knowing this story, you should also lower your expectations for an EM future (a la Hanson and his recent book) and increase your expectations for an AI future in which the AIs are relatively alien.

I recall someone (Grace?) writing a while ago that a surprise AI takeover was unlikely because too many people would see it coming and stop it. That seems less comforting and more whistling-past-the-graveyard now.

Agreed. I always thought EM futures were pretty unlikely, because it required us to be ignorant enough about how intelligence worked that we couldn't modify it significantly but be knowledgeable enough about how it worked to implement it in code (as a simulation, but still).

Is EM, emergent machine or exponentially more or something else?

EM = brain emulation. The idea is that we will achieve AI not by inventing intelligent machines, but by scanning and uploading people's brains.

ornithopter:airplane = EM:AI

I like your way of phrasing it Vaniver. While I think that some in the biz realized that we were coming close, but AlphaGo has really changed the dialogue.

We had a good run but it looks like the sun will be setting soon.

Maybe we can hope Known Space was right about AGI. That would be neat.

For the record, I believe that many people, two years ago, probably already thought that Go had been completely figured out by computers (If you don't believe me, stop reading right now, as you will just feel little more than a condescending desire to disagree with what follows). Within 50 years, AI will be unregulated, just as power is now unregulated, except in the limited circumstances where a localized version of AI supports a seemingly real personality, with the ability to influence other AIs (AIs that cannot influence other AIs but only interact with humans will be unregulated). Yes, AIs *will* , sadly, enjoy regulating each other, based on who their likely designers are... Any AI that can influence another AI will be as highly regulated then as basic TV channels or basic kidney transplants are regulated now, I predict. There was no big news this year, with the exception of a non-zero but non-infinite number of people awakening to the realization that God loves them and that their previous soporific belief that the Bible verses they were familiar with were trivial was a belief that, in its own lovable but superannuated way, was itself an almost trivial belief. (Dear me! as they used to say...) On the other hand, the admirable work put in at correctly parametrizing the multitudinous silicon flip-ups and flip-downs (no rayed vectors here, just x's and o's) that are needed to outnumber the trained choices of lifelong expert GO players was something new and impressive, although not all that news-worthy, due to its incremental nature. Not to say that they do not all deserve however as high a salary as they are getting! By the way, how many people who post on the internet are all that different from the sci-fi characters who, in a slightly different universe, might timelessly wander, in between classes, around the basements of college buildings (in this case, the social sciences building at GMU - in reality, a nice early 70s style 2-story brickish building that smells just a little too much of Lysol, with nicely planted conifers in the lawns around it, where there used to be a spacious sunlit spruce-bordered parking lot to the northeast that was replaced, when enrollment skyrocketed, with a boring multi-story parking garage) looking for bulletin boards with posted statements on it with which they might disagree and to which for some obscure reason they want to reply to, like the soda-intoxicated crowds in old art deco theaters used to hiss at Cagney playing the gangster or used to whistle at Mae West playing the woman who still attracted men long after the lucky ones among their contemporaries had become grandparents? Let's stop mixing up Proverbs 1:1 and Psalms 1:1 (beatus vir qui non sedit in consilio impiorum - five internet related Latinate tags right there).

(describing the building) - with in the past nicely planted conifers near the spacious parking lot in the lawns around it - "skyrocketing enrollments" is, but should not be, another way to say " less trees on campus".

"two years ago, probably already thought that Go had been completely figured out by computers"

Two years ago most experts in the field thought superhuman Go computers were 10-20 years away, and there were enough serious concerns about MCTS scaling that even that wasn't considered a certainty.

This thing moved the state of the art more than 1,000 ELO. Nothing like that ever happened in chess. It's an astounding piece of work.

This was a pretty amazing comment. I'm unconvinced it wasn't actually written by an AI.


I didn't get far into it.

I agree with Lord Action, and I think the 30% figure is wrong by DJF.

I drew, as I'm tired, my gf is calling me, and I can't think straight. Game here: (I should have won this game several times over)

3) "people are even more likely now to believe that men avoid "traditional" female roles"

Step 1: Evading denial of reality.

Many studies (citations needed) have shown that men are, in fact, taking on more household responsibilities traditionally associated with women. However, we are more aware of the remaining gap.

That does seem to be a jarring stance. Anecdotally, I'm surrounded by men who unquestionably take on more inside household responsibilities than their fathers ever did. Most of the articles I've read also indicate men do more. So, this is one stereotype that actually does seem to paint a very poor picture.

"The study authors compared data from 195 college students in 1983 to data from 191 adults in 2014."

It appears this is a typical psychology study where they used local college students as a basis. Which is, of course, a horribly biased sample group. So, the whole study is probably bunk and one that would fail any kind of replication attempt.

Well, the basic logic is that you have to recognize that the issue exists before moving forward on it.

Here's a table on the Canadian situation: Note that females do more housework across all living/martial arrangements. An interesting finding is that even single women do more housework than single men, plausibly because men are less likely to learn cooking skills and might eat out more often, on average.

I think this implies that gender roles continue to explain differences in work in the household and workplace (although I'm not closed to the idea that there may be some natural differences which explain SOME of this).

#7: People move, especially when your neighborhood is going to shit. Come on...

#6 For a moment there I thought Game of Thrones/A Song of Ice and Fire would never be finished. Phew.

An old family friend, son of professors, a phd in economics (naturally), has always been afflicted. For close to forty years, our annual Thanksgiving dinners are a four hour plus endurance run of puns in English, German, Yiddish and Italian. (There may be some in Russian too, but I don't speak the language, and enjoy the absence of assault).

The only virtue is that my lovely spouse ignores the extra glasses of wine and brandy imbibed to counteract the onslaught.

I never looked for a medical definition, but at least now, I have the term, 'Witzelsucht'. Thank you. Oh an undoubtedly there will be a pun to follow when I use it.

#7: Yes of course it did, but an ambivalent-talking, pro-trade acting D still beats an anti-trade-talking R in the rust belt.

#6: Nice one. I think though it misses the huge influence the Beatles had on him. Can anyone imagine the writer of those dull Yellow Submarine instrumentals could have done something as brilliant as the Eleanor Rigby arrangement without that? I guess it's too much to ask that he could have learned to understand and do more with Ringo, whose around-the-beat playing was way ahead of his time. Anyway, if you want to understand the brilliance of George Martin, listen to Let it Be.

"he was the cagey patrician from the metropolis with the posh voice and the fancy education": actually, he was nothing of the sort. His father was a carpenter who lost his job in the depression. The family lived in a small flat, sharing a loo with the neighbours. George Martin attended a local grammar school. Who is Colby Cosh and why does he write such rubbish?

#5 - "And to the extent that policy-makers perceive it to be a problem, some relatively painless solutions exist...investors could, for example, be limited to holdings in only one company per sector, or agree to purchase only non-voting stock, which would leave them with less impact on company management.

Ah yes, we've got to do something about that imaginary stock market scenario that has virtually no possibility of ever happening.

#5 - none of these arguments against indexing seem to have any merit whatsoever. But hey, indexing is hot so let's be contrary.

When there's too much indexing I would expect active investors to start outperforming the index. (Or at least performing competitively with the index). If anything, most comment that alpha is harder to generate than ever.

exactly - I was hoping the article would at least touch on this, but it's just hot air. The interesting question is what % of passive holdings does it take to create new opportunities for active investors. I can imagine it's quite a large %, because even if 50% is held passively there is still plenty left for the price setters.

You are ignoring the fact that active funds, as a group, are also net passive. If "passive" investors own X% of each firm, then the remaining "active" investors, summed up, must own 100-X% of each firm.

(The exception is when stock is issued or retired. So this argument doesn't apply to private equity, for example.)

Good bye George Martin. The Beatles wouldn't have been the same without you (and Brian Epstein; the Beatles started out wanting to be Gene Vincent clones).

My wife said she's leaving me because of my incessant ancient Roman literature puns. I said darling, Aeneid you.

Comments for this post are closed