Thursday assorted links


6. English + Chinese, as in the late and lamented series Firefly.

In my own experience, it was bad English and on trustworthy account worse Chinese still.

Singapore that is.

The problem (or advantage?) of not being able to speak any language properly is much bigger in India. Millions of Indians (if not a billion) have been learning two/three languages for several years.

Amul, the Indian butter company, has been using "Hinglish" in its ads for several years:

This process happens whenever two of more languages are run together-- fairly common throughout history. Sometimes one language emerges dominant, but with strong influences from the other(s). This is the English language today, which is a Germanic language but with a huge amounts of vocabulary from French and Latin.

Bill Easterly's list. Ctrl + F "sachs".

"No matches were found."


I did exactly the same thing.

6) I lived in a former Soviet republic for a while and I thought the merging and mixing of languages was one of the most culturally interesting things about the place. People would start a sentence in Russian and finish it in the local language. And the Russians, who claimed to not know the local language, totally did--such that they could have full conversations where they spoke only Russian and the other person spoke only their local language and they understood each other perfectly well. Beyond that, there were dialects of the local language that let you know what ethnicity the person was. And folks would switch between them, depending on the company. Notably, everyone seemed to know the languages pretty well, so it wasn't a pidgin, like it seems to be in Singapore (Though the couple Singaporeans I've met had 100% command of English). I know the situation is not unique to where I lived, but I think it would be an interesting future of languages if more places end up like Singapore.

1. It's always the 1970s for the monetarists. Unlike the monetarists, Tobin actually moved on from the 1970s and addressed what happened in the 1980s and beyond right up to today: increasing inequality, falling rate of return on capital, stagnant wages, slow economic growth, financial instability. Tobin learned from experience, monetarists never do. When the facts change, I change my mind. What do you do, sir? Tobin's been dead for over 13 years, but his views at the time of his death are more on point than the views on most monetarists today. Monetarist economics can only look good if you're not looking at what happened after the 1970s.

In other words he was wrong and changed his mind?

To the extent Williamson has a point, it seems to be "Yeah, well, Tobin supported wage controls, so what would he know?" Pretty barefaced ad hominem, frankly---not to mention irrelevant to Krugman's point.

(Krugman is full of free policy advice but if he still supports wage controls, that's news to me.)

His point was that Tobin was completely wrong on inflation and Fed commitment.

Just because he drew a picture with lines on it and then reality was also a picture with lines on it doesn't mean he was right.

Unemployment went up to 10%, and NGDP (though not inflation) had inertia. Tobin failed to predict an unexpected positive aggregate supply/productivity shock. There's nothing wrong with that.

Well he failed to predict what inflation would do. It didn't have very high inertia.

Isn't that the whole point?

Nope. It's no intellectual sin to fail to predict positive supply shocks.

#1) Krugman trying to spread his own zombie ideas.

I strenuously object to your execrable libel of our zombie friends.

Plus, contemplate a 20% prime rate or a 12% six-month Treasury Bill rate. Prices/MV of rate sensitive real and financial assets (houses; high-quality, long-term debt securities) fall as interest rates rise, e.g., AAA rated, 6% Telephone bonds pricing at $0.60.

4) “The Cotsbot robot, which has a vision system, is designed to seek out starfish and give them a lethal injection.”

Release the KillBots!


Yes, it is. The "blockchain" is special in that it doesn't require trust to make transactions. If the transaction does require that, then it is just agreeing to keep common records mediated by a third party (to save time, labor, etc.)

None of this means one cannot have a gated, private blockchain. Small scope of the blockchain does not necessarily mean need for trust - as long as there are 2+ parties, consensus may be needed for any change in the ledger.

Yes, a small-scope blockchain does require truest. Worse still, the converse isn't true. The premise behind, say bitcoin, is that it is impractal for a single agent to control more than 50% or more of the mining power.

No one would deny that that idea fails if there are only three miners. Shockingly, it isn't clear whether it work even when there are arbitrary numbers of miners.

The only good I can imagine a "private blockchain" doing would be if you expanded the definition. Cryptographic signatures and hashes provide excellent ways to prove that private agents agreed to some contract, or chain of contract-like documents, even if those documents had originally been a secret.

Egh, half right.

The miner issue is only there to solve the double spending problem, you don't need to rely on the mining pool to solve the verified trader problem. That is 100 percent solved by the signatures alone. So, assuming there is only a single miner, who is trusted to at least show all people the same chain at the same time, you can use a blockchain with arbitrarily low proof-of-work demands on the block hashes, and still get a trust free ledger of ownership.

6. This just occurred to me, but why does Tyler specifically and the blogosphere in general blog so much about Singapore and so little about Luxembourg? In many ways it has an equally remarkable success story (roughly 4 percent post WWII GDP growth until the 70's, just like the rest of Europe, but then decades more growth in the 4-5 percent range right up until 2008, including skipping over several recessions that hit the rest of the continent).

In terms of gdp per capita adjusted for ppp, Luxembourg is the richest country on earth without an obvious "natural resources + small population" explanation for its success, if not the per capita richest nation on earth period. The normal story I hear for that is "it is a seat of European governance so it benefits from wealth transfers from other states." But is that true? Wouldn't that story make us expect shock-driven growth, correlated with years the EEC or EU expanded its scope, not steady year-over-year growth? And if that is the explanation, why isn't Brussels richer?

Someone explain Luxembourg.

Two words: tax haven.

2014 IMF data:

Luxembourg's per capita gdp adjusted for ppp: $92,049
Switzerland's per capita gdp adjusted for ppp: $58,087
San Marino's: $60,664

$30,000+ per capita is a big difference!

Isn't it 500K people surrounded by other relatively well off countries? What is the gdp/capita of San Fransisco?

Exactly...never understood why people ask about Luxembourg or Iceland or even Singapore and wonder why can't Brazil/France/Indonesia/England/whomever be more like them? If a 'country' is really a city's population spread out on a smallish footprint, it's pretty nongeneralizable

According to the Bay Area Council Economic Institute, the Bay Area is the metro region with the highest per capita gdp in the country at $74,815. Still significantly less than Luxembourg! And yet that has been good enough to generate endless articles and blog posts about the Bay Area economy, versus very little speculation about Luxembourg's economy.

In general, if you start googling around for the per capita gdps of various US localities you can get some prestigious suburbs or vacation towns that are higher than Luxembourg, but I'm not sure how that explains Luxembourg, which is neither a prestigious suburb nor a small vacation town.

Luxembourg really is special. It's especially rich for a metro area, for a tax haven, for a center of government, for a micronation, or for anything else it is. I think in a lot of ways it is even more sui generis than Singapore. There's no one cause.

Well there's your explanation, it's sui generis. It's a tiny place full of mostly rich people, and people who make a lot of money helping other rich people avoid taxes who perhaps live elsewhere (so the GDP counts as Luxembourgian). Is it really that hard to figure? Forget SF, what's the GDP/capita of lower Manhattan?

The Bay Area also has 7 million people (broadly defined; not sure I buy those three northernmost counties as truly Bay Area).

In any event, that's an order of magnitude bigger then Lux. There are probably 3-4 subsets you could carve out of the "Bay Area" to end up with 500K populations with significantly higher incomes than $92K.

Manhattan has a Luxembourg level GDP, so that's a good comparison. On the other hand, 1. You can think of Manhattan growth as the whole island gradually becoming as rich as the richest bits, Luxembourg growth appears to be more the whole country getting richer over time. Interesting difference. 2. Manhattan is the heart of a major metro area. Luxembourg is not. "All the rich people in this city live in a rich area" is less surprising than "all the people in this city and in the surrounding countryside are rich." 3. Luxembourg city only contains 1/5 of Luxembourg's population, whereas Manhattan contains all of Manhattan's population. I honestly don't know the extent to which the other 4/5 are living in suburbs and feeding into Luxembourg city vs. the extent to which Luxembourg City is having to subsidize a segment of the population living in a less productive countryside, but it's not a problem Manhattan even has to deal with. I can't find stats for the per capita gdp of just Luxembourg city. 4. Even if Luxembourg is just another Manhattan, how a Manhattan becomes a Manhattan is worth studying and thinking about! Most cities are a Dallas or something equally awful.

Isn't there some issue with people commuting into Luxembourg and living elsewhere?

Luxembourg GDP per capita is quite skewed by people commuting from abroad. Similar story for Liechtenstein and to a lesser extent in Switzerland and Singapore, BTW.

Or just haven.

Singapore is home to the wealth of the rich and well connected throughout the region and beyond. Robert Mugabe is a frequent visitor - his bodyguards roughed up a holidaying Swedish politician who wanted to conduct a citizen arrest in the Kinokuniya bookstore. From the Burmese junta to corrupt Indonesian officials and cautious Russian billionaires, it's the place to be, especially with the crackdowns in Switzerland and elsewhere.

And if you have lots of expensive art that doesn't look to safe in your home country, there are nice duty free warehouses with armed guards.

Low taxes is just half of the story. There are better places to accomplish that goal:

The other half of the story is being in the middle of European Banana of top wealth consumers and having a workforce and laws aimed to excel in logistics.

In Washington DC the per capita Income (GDP) data is massively distorted by the commuters from VA. & Maryland having their output counted in the D.C. production
data but they are not included in the D.C. population.

I've long wondered if something like that was happening in Luxembourg.

I am not sure why Luxembourg or Singapore are considered to be 'nations' in these discussions.

1. They are such small outliers that comparing them with other countries (think USA or India) is outrageous. They are as big as a couple of NY boroughs or a Mumbai suburb.
2. Along of their small size, they do not have the governing complexities of larger economies that have to deal with the checks and balances of various constituencies.
3. They can follow policies (investment, tax, education, social services) that are much more attractive than their bigger neighbors.

Just because they, by some quirk of recent history, have achieved current nationhood does not mean that they are viable long-term states (think 50 or 100 years). They weren't so in the recent past. They sit at kids table, that's all.

If these comparisons are used to get an idea of how different political arangements work economically, then it would be odd to delete examples just becuase of their political arangement.

Maybe the lesson of these places is that all nations should disagragate into networks of large city-states. That idea might not be right, but the scales are tilted if you simply ignore the data.

Being independent since 1839 is not a good test for the viability of Luxembourg? Half (or more) of world's nations are younger. From that perspective, Italy, Canada or Indonesia are quirks of the recent history too.

"And if we think that the world of the future will be largely dominated by the superpowers of America and China, which seems to be the way things have been heading lately, then this kind of English-Chinese bilingualism could become much, much more common. "

Only for Chinese who move to other countries. I can't see anyone else speaking Chinese, not even the children of those migrants.

Yep. English is the default now, the Facebook of languages, garnering all the network effects. You don't have to speak English/be on Facebook but then you are self-isolating.

English has been the default since 1990 or early. That is, until we start speaking binary....

7). I don't get it. There are some inefficiencies in syndicated loan settlements, but it's not just people v computer. In most cases, the borrower and the lead bank both have to consent, not just the buyer and seller. This is because of credit risk ( every syndicated loan has some risk the individual loan holder will have to cough up money for expenses or to disgorge improper payments), reputation risk (terrorist funders), confidential information access risk, and holdout risk in workouts (the last 2 being the big concern to borrower) There are ways to systematize those risks (mostly) and the loan market trade group spends a lot of time doing that, but a block chain seems only to solve a small part of the problem set (keeping track of who took on which risk in respect of a loan that changed hands several times).

7. Yes

People who are optimistic about private blockchains are self-deluded. There is, as of now, exactly 1 blockchain providing meaningful value, namely the bitcoin blockchain (whether core or XT). This realization will set in gradually over the next 2-3 years. 75% chance it eliminates all enthusiasm about blockchains in finance. 25% it eliminates most enthusiasm about blockchains in finance but results in a bigger, badder, more valuable bitcoin. Equilibria are a bitch.

Wow, you obviously have no idea how a blockchain works. Your level or certainty given that fact is impressive.

#2: I remember two things from a course I took in early 2000s that can be described as "ecology for engineers": a) biodiversity is important because it's always good to have redundant life support systems. b) Lions and pandas are not part of Earth's life support systems. That job is done by plants (photosynthesis) and bacteria (decomposers, nitrogen fixation, etc.)

It's nice to see this idea continues being developed, if scientist understand how this work the remediation of polluted sites will advance.

#7. It is not just a database. just because it is not open, does not mean consensus is not required. It is perfectly reasonable to have a much smaller blockchain that is gated and private that still does not require centralization or trust. It will likely be less secure and more vulnerable to collusion than the traditional blockchain but that in itself does not make it a simple "database."

#7 blockchain is a 'proof of work' system - to maintain its security it needs to constantly burn more computing power than any individual entity could reasonably amass and use to disrupt it with the 50% attack. But it is possible to move computing power from mining bitcoin to any other similar system, or from mining litecoin to any other similar system etc. - so the 'private blockchains' will have to burn computing power in rates quite similar to the biggest similar blockchains. Currently bitcoin burns in the environment of a million dollar a day of electricity. This is a crazy cost.

7. Not quite, there are two big differences, one an improvement over traditional databases, and one probably a small detriment.

The improvement is the cryptographic signing done of each transaction. In a traditional database, you have approved agents that authenticate with the database and are allowed to make changes. In a blockchain, anyone can input a transaction, but to be accepted, that transaction must be authenticated by both parties in the contract. This a huge improvement. Before, if you hack an approved agent, you basically now have wide access across many accounts. In the blockchain system, if you hack a particular agent, you have only gained access to make fraudulent transactions on that single agents behalf.

The small downside is that databases are highly optimized for accessing data. This usually takes the form of indexes or partition keys that, much like dewey decimal system, mean that you can find what you are looking for much more quickly than looking through all of the raw data. The blockchain itself doesn't have any such indexes, so if you need to look up a whole bunch of stuff, it could get slow. There's probably some hybrid system that would work here, using the consensus and signing mechanisms to verify new transactions, then putting that information into your own private indexed system for fast lookups.

Comments for this post are closed