Although Stockfish and Komodo have differences in their evaluation scales—happily less pronounced than they were 1 and 2 years ago—they agree that the world’s elite made six times more large errors when on the lower side of equality.
We don’t know how general this phenomenon is, but interestingly it seems to hold much more strongly for top players than for weak players. That is from chess of course.
Here is much more detail from Ken Regan, along with some suggested hypotheses and resolutions.
From Foreign Affairs. Here is my bottom line:
His latest entry into this debate [over stagnation], The Rise and Fall of American Growth, is likely to be the most interesting and important economics book of the year. It provides a splendid analytic take on the potency of past economic growth, which transformed the world from the end of the nineteenth century onward.
In a nutshell, Gordon is probably right about the past, but wrong about the future. My greatest reservation about the work is that Gordon thinks he can predict future rates of productivity growth with considerable accuracy:
…although Gordon focuses on the demographic challenges the United States faces, he never considers that today, thanks to greater political and economic freedom all over the world, more individual geniuses have the potential to contribute to global innovation than ever before.
…many past advances came as complete surprises. Although the advents of automobiles, spaceships, and robots were widely anticipated, few foretold the arrival of x-rays, radio, lasers, superconductors, nuclear energy, quantum mechanics, or transistors. No one knows what the transistor of the future will be, but we should be careful not to infer too much from our own limited imaginations.
And here is the “oops” aspect of the book:
What Gordon neglects to mention, however, is that he is also the author of a 2003 Brookings essay titled “Exploding Productivity Growth,” in which he optimistically predicted that productivity in the United States would grow by 2.2 to 2.8 percent for the next two decades, most likely averaging 2.5 percent a year; he even suggested that a three percent rate was possible.
…Gordon offers a brief history of the evolution of his views on productivity. Yet he does not mention the 2003 essay, nor does he explain why he has changed his mind so dramatically. He also fails to cite other proponents of the stagnation thesis, even though…their work predates his book.
Nonetheless this is a tract well worth reading. Again, here is my entire review.
The author is Sam Bowles and the subtitle is Why Good Incentives are No Substitute for Good Citizens.
Due out in May.
The link is here, the contents include:
A Unit Root in Postwar U.S. Real GDP Still Cannot Be Rejected, and Yes, It Matters: David Cushman examines whether shocks are transitory, permanent, or some of each.
The political ideology of Industrial Relations: Using three metrics, Mitchell Langbert shows the left orientation of the field.
Eli Heckscher’s Ideological Migration Toward Market Liberalism: Benny Carlson explores the intellectual evolution of a great Swedish economist.
Glimpses of Adam Smith: Excerpts from the biography by Ian Simpson Ross.
Classical Liberalism in Econ, by Country: Authors from around the world tell us about their country’s culture of political economy, in particular the vitality of liberalism in the original political sense, historically and currently, with special attention to professional economics as practiced in academia, think tanks, and intellectual networks.
Young Back Choi and Yong Yoon:
Liberalism in Korea
Liberalism in Mexican Economic Thought, Past and Present
(All of the papers from this symposium, which has carried across multiple issues of EJW, are collected at this page.)
2. Henry on John Stuart Mill and liberalism and the Irish famine: “Hence, if progressivism should reasonably be corrected by the Millian tradition on individual liberties, that tradition could do with a lot of correcting back…”
7. More details on Go, note the guy who lost to the program is only #633 in the world. The program is estimated at #279 rank, impressive, but not yet “there.”
Fortune: Hiring a new employee, for instance, now takes 63 days, up from 42 in 2010, according to a 2015 study we did with 400 corporate recruiters. Meanwhile the average time to deliver an office IT project increased by more than a month from 2010 to 2015, and now stands at over 10 months from start to delivery—this particular nugget coming from a study we conducted with 2,000 project managers at more than 60 global organizations.
And when companies need to mesh processes, things get even slower. Multiple surveys we did with several thousand stakeholders in the realm of business-to-business sales revealed some striking evidence of institutional delay. The time required for one company to sell something to another, for example, has risen 22% in the past five years, as gaining consensus from one or two buyers has turned into five or more.
It certainly feels like more people are required to sign off on something than ever before and that fact is slowing things down. The time-series is short, however, and lots of other things are going on. Maybe firms take longer to hire when the growth rate is low. File under speculative.
Mexican non-oil exports to USA in December (y/y): -4.5%. Excluding autos: -8.7%.
It’s funny how these numbers seem to indicate someone is starting to enter a recession. Who might that be? Maybe it’s just noise, I don’t see any other mediocre economic reports wandering around these parts… Or maybe it’s Mexico that’s the problem…
The excellent Susan Athey addresses that question on Quora, here is one excerpt:
Machine learning is a broad term; I’m going to use it fairly narrowly here. Within machine learning, there are two branches, supervised and unsupervised machine learning. Supervised machine learning typically entails using a set of “features” or “covariates” (x’s) to predict an outcome (y). There are a variety of ML methods, such as LASSO (see Victor Chernozhukov (MIT) and coauthors who have brought this into economics), random forest, regression trees, support vector machines, etc. One common feature of many ML methods is that they use cross-validation to select model complexity; that is, they repeatedly estimate a model on part of the data and then test it on another part, and they find the “complexity penalty term” that fits the data best in terms of mean-squared error of the prediction (the squared difference between the model prediction and the actual outcome). In much of cross-sectional econometrics, the tradition has been that the researcher specifies one model and then checks “robustness” by looking at 2 or 3 alternatives. I believe that regularization and systematic model selection will become a standard part of empirical practice in economics as we more frequently encounter datasets with many covariates, and also as we see the advantages of being systematic about model selection.
…in general ML prediction models are built on a premise that is fundamentally at odds with a lot of social science work on causal inference. The foundation of supervised ML methods is that model selection (cross-validation) is carried out to optimize goodness of fit on a test sample. A model is good if and only if it predicts well. Yet, a cornerstone of introductory econometrics is that prediction is not causal inference, and indeed a classic economic example is that in many economic datasets, price and quantity are positively correlated. Firms set prices higher in high-income cities where consumers buy more; they raise prices in anticipation of times of peak demand. A large body of econometric research seeks to REDUCE the goodness of fit of a model in order to estimate the causal effect of, say, changing prices. If prices and quantities are positively correlated in the data, any model that estimates the true causal effect (quantity goes down if you change price) will not do as good a job fitting the data. The place where the econometric model with a causal estimate would do better is at fitting what happens if the firm actually changes prices at a given point in time—at doing counterfactual predictions when the world changes. Techniques like instrumental variables seek to use only some of the information that is in the data – the “clean” or “exogenous” or “experiment-like” variation in price—sacrificing predictive accuracy in the current environment to learn about a more fundamental relationship that will help make decisions about changing price. This type of model has not received almost any attention in ML.
The answer is interesting, though difficult, throughout. Here are various Susan Athey writings, on machine learning. Here are other Susan Athey answers on Quora, recommended. Here is her answer on whether machine learning is “just prediction.”
Bloomberg: Apple Inc. said it acquired education-technology startup LearnSprout, which creates software for schools and teachers to track students’ performance.
Apple is working on education tools for the iPad, which will allow students to see interactive lessons, track their progress, and share tablet computers with peers….More than 2,500 school districts in 42 U.S. states use LearnSprout’s software, according to the company’s website.
As I said in my post, Apple Should Buy a University:
Apple University would be a proving ground for educational technologies that would be sold to every other university in the world. New textbooks built for the iPad and its successors would greatly increase the demand for iPads. Apple-designed courses built using online technologies, a.i. tutors, and virtual reality experimental worlds could become the leading form of education worldwide. Big data analytics from Apple University textbooks and courses would lead to new and better ways of teaching. As a new university, Apple could experiment with new ways of organizing degrees and departments and certifying knowledge.
Andrew Batson thinks it is simpler than many people make it out to be:
…these analyses…fail to even mention the most straightforward and direct explanation of why China’s growth is much slower today than it was in say, 2010 or 2007. It’s not like it’s a secret. From about 2003 to about 2010 China had the biggest construction boom of modern times and probably in all of human history. Then in 2011-12 the construction boom ended. That’s it. Really, that’s all you need to know. Well, you might need one more fact: housing and construction account for as much of a third of China’s GDP, once all their indirect linkages to other sectors are considered. I think a housing downturn explains very well the timing, severity and distribution of the economic slowdown that has actually occurred.
Here is the full post, which also criticizes the idea of the middle income trap. I would add two points, which may represent a deviation from Batson’s argument. First, I don’t think the Chinese growth slowdown is as sudden as a culling of media reports might suggest. Second, to the extent the contraction is sudden, it is perhaps Chinese investors have woken up to the idea of a risk premium, and realized there is no eternal ten or even seven percent growth to validate so-so quality investments. The dynamics of information arrival can compress economic adjustments into “too short” a space, a common theme in business cycle theory and not an issue restricted to contemporary China.
6. Move to Florida, wake up with a kinkajou on your chest.
Tic-tac-toe fell in 1952, checkers in 1994, chess in 1997 and it now looks like Go, the ancient Chinese game that has a search space many, many times greater than chess, has fallen to a new AI from Google.
…our program AlphaGo achieved a 99.8% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0. This is the first time that a computer program has defeated a human professional player in the full-sized game of Go, a feat previously thought to be at least a decade away.
Importantly, AlphaGo isn’t based primarily on searching a huge space but on deep neural networks that learned first from human players and then from simulated play with itself. The techniques, therefore, are not limited to Go.
AlphaGo will face its greatest challenge in March.
AlphaGo’s next challenge will be to play the top Go player in the world over the last decade, Lee Sedol. The match will take place this March in Seoul, South Korea.
Win or lose, I will bet that Lee Sedol is the last human champion the world will ever know.
The subtitle of Thomas Leonard’s new and excellent book is the apt Race, Eugenics & American Economics in the Progressive Era.
I take it you all know by now this is quite an ugly story, namely that both early progressives and late 19th century American economists were often quite appalling racists and eugenicists, and that such racism was built into the professional structure of economics in a fairly fundamental way, including but not restricted to the American Economics Association.
Kevin Drum had an interesting point in response (and do read his full post, there is more to it than this quick excerpt):
Early 20th century progressives supported eugenics out of a belief that it would improve society. Contemporary liberals support abortion rights and right-to-die laws out of a belief in individual rights that flowered in the 60s.
Most of all Drum is saying that the earlier history is not very illustrative of anything for today.
I view it this way. Go back to Millian liberalism of the mid-19th century. Had American or for that matter British Progressivism been infused with more of this philosophy, the eugenics debacle never would have happened. For instance if you look at the British Parliamentary debates of 1912 over the Mental Deficiency Bill, the anti-eugenics forces drew heavily upon Mill for their inspiration. This was standard stuff, but the Progressives of the time didn’t see much of a pro-liberty reason for being pushed into a Millian position, quite the contrary.
The claim is not that current Progressives are evil or racist, but rather they still don’t have nearly enough Mill in their thought, and not nearly enough emphasis on individual liberty. Their continuing choice of label seems to indicate they are not much bothered by that, or maybe not even fully aware of that. They probably admire Mill’s more practical reform progressivism quite strongly, or would if they gave it more thought, but they don’t seem to relate to the broader philosophy of individual liberty as it surfaced in the philosophy of Mill and others. That’s a big, big drawback and the longer history of Progressivism and eugenics is perhaps the simplest and most vivid way to illuminate the point. This is one reason why the commitment of the current Left to free speech just isn’t very strong.
I don’t mean to pick on Kevin, who is one of my favorite bloggers, but I disagree (and find indicative) another one of his claims, namely:
…eugenics died an unmourned death nearly a century ago.
To give one (not the only) example to the contrary, Swedish “progressive” sterilization persisted through the 1970s, as was true for Canada as well. Eugenicist views toward autistic people, among others, remain common across the political spectrum (no special brickbat for Progressives here, but they are guilty too), and with CRISPR a lot of eugenicist debates are already making a comeback.
Do we really want to identify with a general philosophy which embraced eugenics for so many decades, when so many pro-liberty and also social democratic thinkers were in opposition? I think Mill himself would say no.
From my inbox, from Bruce Caldwell:
The Center for the History of Political Economy at Duke University will be hosting another Summer Institute on the History of Economics this summer, May 29-June 17. The three week program is sponsored by the National Endowment for the Humanities and is designed primarily for faculty members in economics, other social sciences, and the humanities, though three of the twenty-five slots are reserved for graduate students. Participants will be competitively selected and successful applicants will receive a $2700 stipend for attending, out of which they will pay for their own room and board. Our line-up of discussion leaders is quite impressive, and includes Maria Pia Paganelli, Nicholas Phillipson (author of Adam Smith: An Enlightened Life), Bart Wilson, Duncan Foley, Tim Leonard, Angus Burgin, Eddie Nik-Khah, and Steve Medema. The deadline for applying is March 1. A special bonus for those who attend: the History of Economics Society meetings will be held at Duke from June 17-20. Attendees who wish to do so can stay over for the HES meetings.
More information on the Summer Institute is available at our website,