Tuesday assorted links

1. Luigi Zingales on the Laureates.  And the full 49-page report on the Laureates, from Sweden.  And David Warsh on Holmström on debt and bankingCardiff Garcia on Bengt on moneyKling on the prizeBoettke on the prize.  A Fine Theorem on Oliver Hart, and how he changed his mind, recommended.  And where Nobel Laureates come from.

2. The economics of Twitter.

3. Tim Harford on man-machine interaction, adopted from his new book Messy.

4. College chain from India expanding into the U.S.?

5. Paul Krugman on Brexit and the pound. I don’t agree with those arguments, but I think they are more consistently Krugmanian than his earlier posts on Brexit.  If an economy is in demand-side secular stagnation, and then a government taxes production and trade, and takes finance down a peg, but boosts exports through a lower currency, I am not sure the model will predict significantly negative consequences.  His arguments are also consistent with a more general New Old Keynesian neglect of wealth effects.

6. Why does the UK have such a radical Brexit strategy?

Comments

You do not need much of a model too forecast Brexit impact.
The negatives work on investment over the long run with big lags.
A weak currency encourages more production with a relatively short lag.

So you should expect the first impact of Brexit to be a boost to growth in the short run.

How does a hard Brexit affect Scottish independence? It may make them want independence more, but surely England and Wales are still the natural largest trading partner for an independent Scotland. An independent Scotland in the EU with a free trade deal and free movement with England would be a backdoor soft Brexit (especially if Scotland signed up to Schengan) that the EU seems ill inclined to grant, but a Scotland lacking such would have a pretty severe negative impact to its economy.

That's about right. Scotland does about four times as much trade with the rest of the UK as it does with the entire rest of the EU. In addition, Scotland currently runs about a 9% of GDP annual fiscal deficit, part of which is made up by fiscal transfers from the rest of the UK. So Scotland doesn't really have any easy options. Any change to their current situation costs them a lot.

So is there a cynical argument that hard Brexit makes Scottish independence nigh impossible, and some in the EU or Rest of UK push it for that reason?

the story I have heard is that formally the UK is the member and if it leaves, then Scotland would need to re-apply. the other EU states with separatists (e.g. Spain and ETA) will then be loathe to make a special case for Scotland joining the EU without the UK, pour encourager les autre.

A totally unnecessary headache.

Instead of removing Scotland and leaving the EU, just have England withdraw from the UK. The Untied Kingdom of Scotland and Northern Ireland is already an EU member and doesn't have to do anything. England is out of the EU without any touchy formalities. Everyone is happy.

ETA is no more and Basque separatism is at its lowest point since polls on the issue exist. It is in Catalonia where the problem lies.

Imagine the irony if England responded to another Scottish independence push with a hard Brexit-type of response in which all deals would start from scratch.

A bitter irony for the pro-EU Scots, and particularly bitter because unlike England and the EU, the Scottish get a lot more from union than without.

All deals would certainly start from scratch, as there are no existing Anglo-scottish deals to base anything on.

There was a little bit of that in the recent referendum. The pro independence Scots insisted, "hey, of course we can continue to just use / have parity with the British pound, look at all the little Caribbean and Central American countries who have unilaterally dollarized" while the BoE tried to insist that somehow it wasn't that easy.

There was, wasn't there? But it was to my mind a bit more like Mark Carney trying to logically dispute why it might be better for Scotland not to be allowed to use the pound, unless they compromised on sovereignty. Fairly logical arguments (even if you didn't agree with them) rooted in "We shouldn't do this, because..."

Whereas with Brexit, at least by the time it hits the British press (where perhaps it may have been mangled quite badly), there isn't much left of logical arguments from the EU about why deals can't work than "Everyone would want an a la carte Europe and we must dissuade potential -exiters". There's not much "We wouldn't want a non-EU country in the Single Market because that's inherently a bad idea" logical argument, just an impression like "If you're in the club, you follow the rules of the club, and if you're out of the club, you're out of the club!" and "The Four Freedoms are indivisible" with no logical or economic argument to defend that.

The Scotland Referendum didn't, notably, revolve around the idea that in the event of a Scottish exit the rUK must treat an exiting Scotland harshly, and utilize all the strength of its greater economy without restraint, in order to avoid Wales exiting the Union, or a separate kingdom of Yorkshire or city state of London emerging, or anything like that. Nor fear of populist SNP wannabe separatist parties springing up all over the UK. Nor justification that the UK must continue to exist lest we return to the age of war between Scotland and England.

There was also some (entirely reasonable) questioning about 'what are you going to do about all the English living in Scotland (and vice versa)?" but I don't think either side talked about tossing them out the way is happening now. I suppose there's the question of what would happen after a Yes, though.

Then again, the No side and Rest of UK could just lean on the fiscal transfers from England to Scotland, whereas with the net flows in the other direction for UK/EU, it appears that the EU wanted to play hardball.

1. Where Nobel laureates come from :

Often from other countries

http://www.huffingtonpost.com/james-witte/nobel-laureates_b_2458128.html

the most interesting thing probably is how many of them are Jews (both the recent British American winners for example). Not sure that gives such a neat little immigration story though.

1. Where do criminals come from :

Often from other countries

https://en.wikipedia.org/wiki/Shooting_of_Kathryn_Steinle

#2 Seems to be based around the assumption that defeating trolls/spammers and running a web service of the magnitude of twitter are all trivially easy tasks and therefore, the 1.5 Billion that Twitter spends on R&D must be going toward hookers and blow or something.

I have an alternate hypothesis, those things are not easy and therefore they are expensive to do right.

It seems the problem is that each time they shut down a troll their traffic goes down.

This is like trying to clean up CB radio by kicking the truckers off.

Pfft. Twitter has leveled off but never lost active users.

https://www.statista.com/statistics/282087/number-of-monthly-active-twitter-users/

(Unless you count a little 307M to 305M noise as signal.)

That doesnt really tell you what the ratio of real users to fake ones is, though. They could be losing real users and gaining fake ones at about the same rate.

The simplest assumption is a similar ratio over time. Certainly we can exclude an explosion of bots because that would be an explosion of accounts. It would also be too much to assume that an increase in bots would perfectly balance a fall off in humans. So ... probably as it looks, everything found stasis.

Stasis does imply that twitter is having no real luck in finding and removing bots/spammers. Not saying your wrong, just pointing that out.

There is a good and a bad side to the open API. "How to create a Twitter bot" is a kid's lesson. And a spammer's paradise.

Let me see if I can find a "good" bot ..

https://twitter.com/TallikaRoan

Boring, but we all know the temperature of Roan's room.

You don't need $806 million a year in R&D to fight spam.

anon- I disagree about assuming the same rate. Bots, etc., are increasing in number and ability, so if there is an assumption to be made about trend, it is positive and in that direction. However, for practical purposes of numerical estimates often it's better to formally assume zero change (bots and sockpuppet accounts to human ratio stays the same), and then turn around and say you know it's positive, but would rather say zero than take a stab at which positive number to plug into the analysis.

Suggesting that the bot growth rate is positive and that the ratio is constant is actually numerically possible. The leveled off, slow, growth at Twitter could be humans and bots a the same old ratio.

And just to go back to why there would be humans, if you know how to use it, it can be a relatively spam-free news source.

The fastest news source on the planet.

If you follow trusted people, it won't even have many false positives. For MR type things try these feeds:

@TylerCowen of course, and @amcafee @bencasselman @BEworksInc @b_fung @codytfenwick @davidmanheim @DouthatNYT @erikbryn @esoltas @ezraklein @felixsalmon @foxjust @JustinWolfers @KevinJBoudreau @LHSummers @matt_levine @mikerosenwald @NateSilver538 @Noahpinion @Nudgeblog @planetmoney @PolitiFact @prchovanec @PTetlock @ramez @ReformedBroker @robotenomics @R_Thaler @samirvarma @TimDuy @willwilkinson

I'm talking about relative growth, not absolute growth. Given technological advances, it will be relatively more used, and grow faster than the number of people coming online. It's not rock solid logic for 100% of all possible cases, but consider where betting markets would end up on a question like this. I rather doubt you'd even get any takers on the other side, once money were on the table.

He probably is going short on Twitter .......

http://www.bloomberg.com/news/articles/2016-07-11/the-fraud-hunting-short-seller-taunting-bill-ackman-on-valeant

Yeah, in theory in algorithm used to detect spammers could be used by the spammers to avoid such detection. They could probably do more, but for all we know the revenue is all ads by penny stock promoters or people who want to sell to them anyway. Plus, it starts to get into censorship; what if someone really wants to talk about penny stocks?

I never could get into twitter even when they offered me a job there. I think it will be tough for them to fire most of their staff and keep the rest, if they fire everyone they wont be able to stay efficient as web and mobile technologies change. It is hard to see a financial buyer, I would want a lot of collateral to lend to them.

Twitter basically took the "give it away, make it up in volume" thing too far. Their page is too sparse. Their API is too open. It would be sad to see it go, just because no one else will make that "mistake".

As an aside, today is a pretty amazing day on Twitter. Trump is self-immolating and followers are deciding whether to throw themselves on the pyre.

https://twitter.com/realDonaldTrump/status/785854588654092290

A bit of context: Facebook, G-Mail, Hotmail or whatever spam control work much better than Twitter. So, is that hard?

Two points, 1) How much do those guys spend doing so and b)Filtering spam email is different from filtering spam twitter. If nothing else, you can filter based on the sending email server. Twitter is kinda designed to be one to many, spammy from the start, so its a bit more challenging to filter. Just a thought on that last one, i dont do filtering or twitter so im just throwing stuff out there.

Is spam the problem, or is it trolling? Because only one of those seems possible to combat algorithmically, and that's the one I see complained of most often.

Trolls are also easy to deal with by a simple rule like "if 5 registered users click ABUSE then you're banned for 24, 48, then infinite hours", but there's a risk of leading towards echo chambers.

As a centrist who on occasion gets booted from extremist sides on both ends by popular demand, I can testify to the existence of this phenomenon.

Perhaps it's not that hard. Instagram it's also has "spammy" potential but regularly culls bot profiles.

Looking at this plot, it makes sense for Twitter to be soft with the bots. Twitter user base stopped growing on Q1 2015......18 months ago. Just imagine this plot if they went hard on the bots.The CEO would have a hard time to explain how costs go up but user base goes down. https://www.statista.com/statistics/282087/number-of-monthly-active-twitter-users/

It seems Mr. Hempton missed the active user statistics. How operating expenses can go up while user base is stagnant?

Block anyone who spams/harasses you and don't follow people with low average quality of shares. What's the spam problem on Twitter? I haven't used it enormously, but I never ran into any problems.

Gmail is great for spam - just about any "cold calling" out-of-the blue spam that has made it into the inbox in recent years was nominally a legitimate offer (e.g., "I see you have a website, now do you want to buy web design services?")

Facebook has a similar risk as Twitter, in terms of possibly following multiple accounts of people for reasons you don't clearly remember, but who feed targeted garbage into your newsfeed. But mostly I follow think tanks, and individuals involved in various advocacy activities, on Facebook, and generally speaking it's pretty easy to identify the trolls, spammers and psy-ops folks by looking through their feed, or at worst within a couple/few days they'll reveal themselves and you can either eliminate the influences from the newsfeed or just observe which kinds of misinformation are being presented and how it might be intended to impact you.

I'm not a twitter fan. I don't care either way. However, the author seems to be playing loose with the obvious facts.

"Revenue has gone up very nicely - from $664 million to $2.2 billion and is still increasing. And costs have gone up commensurately. Losses seem stubbornly stuck at half a billion per annum. "

Then looking at the chart directly above the comment: 2014 ($647); 2015 ($578); 2016 ($533).

Those losses aren't stubbornly stuck at half a billion. He's crafting a narrative and not even bothering with anymore than a tenuous connection with reality.

They haven't made much progress in defeating spammers, the functionality of twitter is virtually unchanged, and operating costs aren't going to be counted as R&D. So the question is, what did that money buy? The most likely answer to that question is "nothing."

4) I really don't think Boards of Education should be able to block sales of "for-profits art colleges." Frankly I don't know why for-profit art is considered Education in that sense. Pure frippery and not related to state welfare.

Almost every article I read from the Guardian is an illogical mess.

"Crash: how computers are setting us up for disaster ... This makes it very hard to crash an A330, and the plane had a superb safety record: there had been no crashes in commercial service in the first 15 years after it was introduced in 1994. But, paradoxically, there is a risk to building a plane that protects pilots so assiduously from even the tiniest error. It means that when something challenging does occur, the pilots will have very little experience to draw on as they try to meet that challenge."

Ummm, ok, we have an assiduous auto-pilot that results in a superb safety record.

It's clear that automation interfaces should give clear signals to operators. That's a key stone of the design process, but it's often buried under conflicting demands. However, as the facts indicate with respect to modern planes, the results have been worthwhile.

Planes are an especially odd example to choose for illustrating the risks of this phenomenon, because there's a pretty significant limit to the worst-case outcome--the plane crashes and everyone dies. The only question is whether that happens more often or less often with the autopilot (with the answer being less often).

Compare that to automated financial trading or automated medical diagnostics, where the consequences can end up being exacerbated by technical malfunctions. In those fields it is perhaps a source for some circumspection in deployment. In planes? Not so.

The article is just high tech Ludditism.

"However, the younger generation of meteorologists are happier to trust the computers."

"Anuj Pradhan has floated the idea that humans should have to acquire several years of manual experience before they are allowed to supervise an autonomous car. But it is hard to see how this solves the problem. No matter how many years of experience a driver has, his or her skills will slowly erode if he or she lets the computer take over. "

Well I haven't raised a food crop since I was a teenager. I guess we should all get back to the basics of subsistence farming. How can we trust all these plow horses? You need at least a few years of hoeing the ground by hand to be able to supervise an autonomous plow horse.

And don't get me started about all these youngsters and their tractors. Have you thought about what you are going to do when the fuel runs out?

It's a low-brow echo of Charles Perrow. He was wrong too. Computers, and technology in general, make risk easier to manage and lower the probability of disaster.

Perrow thought we were on some trajectory of ever increasing disaster frequency and severity as increased complexity made things harder and harder to manage, and the opposite has been the case.

"The article is just high tech Ludditism."

+1

Par for the course for modern liberalism.

The errors of reasoning they point out *are* noteworthy. But the solution isn't that, at the end of the day, computers are gonna compute and you just need a good 'ol human brain to fix errors. The solution is a next-gen of risk mgmt software/ML Algos (some call it AI, but that's such a buzzy word). The author makes the classic error of not realizing our brain is just a specific case of the general computer.
I suspect in 15-40 years the idea that a *human* was controlling a plane will be comically terrifying and a sign of a technologically primitive past.

I had the same feeling on reading this article that there was a whole bunch of lazy reasoning and sensationalism and speculation. This is very unusually poor work by Hardford, usually he is interesting and thoughtful. I wonder if he is writing too much now?

That automation has made flying safer seems obvious; nonetheless, it should also be obvious that unused skills- such as "hand flying" an airplane- will indeed get rusty when not used, and in an emergency may not be available to meet the challenge.

Then again, the solution is almost equally obvious: require pilots to regularly train in simulators, where emergencies can be simulated which no one would dare deliberately create with a real airplane and real lives at stake. Simulation is never quite the real thing (if for no other reason than because one knows it's a simulation), yet the same technology that enables flight automation has also made it possible to build very good simulators.

That the Guardian does not mention this obvious risk mitigation says little good for the Guardian; nonetheless, the risks of over-reliance on automation seem real enough.

So, should Navy and merchant marine academies require would-be ocean navigators to learn celestial navigation against the day when GPS fails?

"So, should Navy and merchant marine academies require would-be ocean navigators to learn celestial navigation against the day when GPS fails?"

Well yes and indeed they have brought back the requirement.

"U.S. Navy Brings Back Navigation By The Stars For Officers"

http://www.npr.org/2016/02/22/467210492/u-s-navy-brings-back-navigation-by-the-stars-for-officers

I'm not sure whether that's useless credentialism, because they need to keep the officers busy, or a really good idea, because in the military you face an intelligent adversary who could actually take your systems out.

Pretty clearly the latter. Forget cyberwarfare, merely blocking military forces from their navigational sensors renders them blind. And those GPS satellite signals are easy to jam (and a lot of commercial navigational units use the Russian GLONASS satellites, not exactly a reliable source of information if hostilities should arise).

It's still useful to learn Morse code too, as US POWs learned in Vietnam.

Jim Lovell used a sextant for navigation back to Earth on Apollo 13.

I think outlet fisks are the best fisks. Very current as well. I heard from some orange guy on Twitter that CNN is a bad outlet and we shouldn't listen.

anon, your comments are degenerating into an incoherent mess.

Surely that was accessible. Tyler introduces a piece with a byline. Not only that, with the byline of an author oft cited and discussed at MR.

But let's go off on the outlet instead: "Almost every article I read from the Guardian is an illogical mess."

You are making a much bigger (and sillier) claim than "Tim Harford is wrong on this one."

As an aside, it is possible Tim Harford got a little over-excited in his man bites dog story, but only a little, and not enough to condemn an entire outlet.

"Surely that was accessible. "

No, the comment you made was not accessible. It wasn't even intelligible.

Here is what you wrote:

"I think outlet fisks are the best fisks. Very current as well. I heard from some orange guy on Twitter that CNN is a bad outlet and we shouldn’t listen."

Frankly, that's one of the most rambling, incoherent comments I ever seen on this blog.

I can explain it more, if it would make you feel better (or worse).

It is now a standard practice for extremists, on the right and the left, to say "oh no, don't trust that outlet." The orange man does it:

https://twitter.com/realDonaldTrump/status/785563318652178432

Now why would they do it? Of course it is because striking off an outlet is even better than arguing a fact or a logical progression. If you can succeed you never have to consider CNN (or the Guardian) again. You have completely rejected critical reading from there on out. All you need is the very most superficial check possible.

Who wrote it? The Guardian? CNN?

Well then gosh, we don't need to pay attention at all.

(I give you a hard time, but Alain's laying all of " modern liberalism" at Halford's feet was probably a bit worse.)

So, let me get this straight. I wrote a critical comment about a British tabloid, involving an article on technology with one rather obvious critical logical flaw and a general Luddite tone. A comment whereby I specifically laid out the flaw.

The article and my responses have absolutely nothing to do with American politics. And yet you responded by jumping to some obscure Trump bashing. At least from your followup I assume that the comment "some orange guy on Twitter" refers to Trump.

"It is now a standard practice for extremists ... The orange man does it:"

I guess in your mind, referring to Trump as an "orange man" and attempting to establish a connection between Trump and any completely unrelated comment you disagree with, is a normal reaction, but to me it indicates that you are one of those extremists you deride.

You probably need to look up "tabloid," but that isn't terribly far off.

Your lead was "don't trust the Guardian," a British national daily newspaper founded in 1851. Time for a blurb:

In 2008, Guardian columnist Jackie Ashley said that editorial contributors were a mix of "right-of-centre libertarians, greens, Blairites, Brownites, Labourite but less enthusiastic Brownites, etc," and that the newspaper was "clearly left of centre and vaguely progressive". She also said that "you can be absolutely certain that come the next general election, The Guardian's stance will not be dictated by the editor, still less any foreign proprietor (it helps that there isn't one) but will be the result of vigorous debate within the paper."

Do you honestly think it was editorial quality that made you lash at the Guardian, rather than discuss Harford and his ideas?

#5: Brexit has not happened yet. No one knows which will be the trade rules on 2019. Whatever change seen today is driven by expectations and educated guesses. If expectations have an effect, it will be seen on the Q3 2016 GDP report by the Bank of England coming in 2 weeks. Confirmation comes on Q4 2016 report by the end of January 2016. Krugman should know the commonly accepted definition of recession is two quarters of negative growth. He can say no major business has gone bankrupt and unemployment rate is stable since March 2016 but telling recession or not is an issue of the BOE. Krugman also says a cheaper GBP is good for manufacturing exports. This is true as long as the GBP depreciation is larger than new tariffs applied to UK products after Brexit. If not, the depreciation is just an illusion.

I will bet you $100 that the UK will not go into recession this year or next. They only way this would happen is if the BOE got scared of inflation and seriously hiked interest rates. But the BOE have already said that they are willing to tolerate a little extra inflation (they get it). The recent fall in the pound is all about the BOE now adapting a much looser monetary stance compared with the ECB and the FED both of which are saying they will raise interest rates as soon as they can, and the BOE saying that they are considering lowering interest rates. This is going to be a great test of how economy respond to domestic demand stimulation.

Perhaps you are also correct that the EU will in the long term apply tariffs on UK exports, but if these are applied by the EU it actually is a very good example of why the UK should not be in the EU. Raising tariffs is a mercantile approach, you think you are punishing your trading partner but you are actually punishing yourself. No country should want to part of a mercantalist block, it is all about making the state strong at the expense of the population. If I am threatened by someone who says they will harm themselves unless I do what they want, I don't regard that as 1) a very powerful threat and 2) I am less likely to agree to join them in a business venture as a result, even if I might lose some business with them. In the really long run, better to stay away from such people.

Raising barriers, both tariff and non-tariff, is a bit relative term. From today's point of view tariffs will rise. But from an historic point of view is just a back to non-free trade agreement levels.

It's like your ex refusing to have sex with you after breakup or divorce after you ask. From the relationship point of view it looks like your ex is imposing barriers, withholding something, punishing you and punishing herself.. But from a more general point of view, not having sex it's just the average relationship between women and men.

Raising tariffs is more like you trying to stop your roommate from having sex with someone you don't like by charging him money for the privilege. The person he is trying to have sex will go off and have sex with someone else leaving them no worse off but your friend is worse off and you have nothing but a power trip (you started this metaphor....). The EU functionary stopping a French person who wants to buy a British car because they are pissed off at the British for leaving their club is perhaps a better illustration.

#6: They simply refuse to acknowledge that minor changes in the effective terms of trade will have minor effects, and are at pains to pound the table to avoid telling their readers what they truly object to.

#1 - A Fine Theorem says: "So where does all this leave us concerning the initial problem of why firms exist in a sea of decentralized markets? In my view, we have many clever ideas, but still do not have the perfect theory. A perfect theory of the firm would need to be able to explain why firms are the size they are, why they own what they do, why they are organized as they are, why they persist over time, and why interfirm incentives look the way they do" - seems overly ambitious.

This whole field reminds me of evolutionary biology--full of fine theories that are essentially untestable. A load of rubbish.

#1, where Nobel Laureates come from: How much of this is an artefact of the data? Or alternatively, just a good selection criteria to find the smartest people in the world in that year?

If I have few students, then my 'nobel laureate hit rate' might be very high, at least for a few years anyway. I don't see that this is any kind of useful metric to say that these institutions are necessarily doing an excellent job at value adding to their students when they arrive. Alternatively, perhaps networking has benefits in who you know as well, and nobel prizes are handed to people that you know.

If you don't study under a Nobel winner, then how are you supposed to get the attention to get the resources to get the people and data to do the big research? Surely there are many many millions of people on the planet perfectly well-suited to earning a Nobel in one thing or another, but who will never meet the circumstances that will enable them to get themselves there.

1: sources of Nobel prizes. I liked Hsu and Wai's earlier article better; they looked only at science achievements and only at US colleges and universities, but this meant that they could use an achievement measure with larger sample sizes hence greater reliability: alumni who've been elected to the various national academies (of science, engineering, and medicine).

Small liberal arts colleges are even more numerous on this list of top producers.

http://qz.com/498534/these-25-schools-are-responsible-for-the-greatest-advances-in-science/

Comments for this post are closed