Results for “solow” 86 found
An aggregate Bayesian approach to more (artificial) intelligence?
It is not disputed that current AI is bringing more intelligence into the world, with more to follow yet. Of course not everyone believes that augmentation is a good thing, or will be a good thing if we remain on our current path.
To continue in aggregative terms, if you think “more intelligence” will be bad for humanity, which of the following views might you also hold?:
1. More stupidity will be good for humanity.
2. More cheap energy will be bad for humanity.
3. More land will be bad for humanity.
4. More people (“N”) will be bad for humanity.
5. More capital (“K”) will be bad for humanity.
6. More innovation (the Solow residual, the non-AI part) will be bad for humanity.
Interestingly, while there are many critics of generative AI, few defend the apparent converse about more stupidity, namely #1, that we should prefer it.
I am much more worried about #2 — more cheap energy — than I am about more generative AI.
I don’t know anyone worried about “too much land.” Maybe the dolphins?
There have been many people in the past worried about #4 — too many people — but world population will be shrinking soon enough, so that is a moot point.
I do not hear that “more capital” will be bad for humanity. As for innovation, the biggest innovation worriers seem to be the AI worriers, which brings us back to the original topic.
My general view is that if you are worried that more intelligence in the world will bring terrible outcomes, you should be at least as worried about too much cheap energy. What exactly then is it you should want more of?
More land? Maybe we should pave over more ocean, as the Netherlands has done, but check AI and cheap energy, which in turn ends up meaning limiting most subsequent innovation, doesn’t it?
If I don’t worry more about that scenario, it is only because I think it isn’t very likely.
If you worry about bringing too much intelligence into the world, I think you have to be a pretty big pessimist no matter what happens with AI. How many other feasible augmentations can have positive social marginal products if intelligence does not?
Addendum: I have taken an aggregative approach. You might think we need “more intelligence” and also “more AI,” but perhaps in different hands or at different times. In contrast, I think we are remarkably fortunate to be facing the particular combination of parties and opportunities that stand before us today.
Monday assorted links
1. Sensible talk from the Left about how little the Left is offering so many American men.
2. Where did our belief in abundance come from? The Bible, no?
4. milky eggs on life extension. I don’t endorse (or deny) any of that, but why not more discussion on pursuing higher social status?
5. U.S. per capita carbon emissions, over time.
6. Why did Oppenheimer end up on St. John for the last years of his life?
Friday assorted links
1. Tymofiy Mylovanov on foreign aid in Ukraine.
2. The most beautiful post offices? Why are none of them recently built? Recommended.
3. The contrarian case for Pakistan’s upside.
4. AI girlfriends markets in everything. And LLM liability for financial advising — what is the right approach?
5. Florida insurance carrier update.
A Big and Embarrassing Challenge to DSGE Models
Dynamic stochastic general equilibrium (DSGE) models are the leading models in macroeconomics. The earlier DSGE models were Real Business Cycle models and they were criticized by Keynesian economists like Solow, Summers and Krugman because of their non-Keynesian assumptions and conclusions but as DSGE models incorporated more and more Keynesian elements this critique began to lose its bite and many young macroeconomists began to feel that the old guard just weren’t up to the new techniques. Critiques of the assumptions remain but the typical answer has been to change assumption and incorporate more realistic institutions into the model. Thus, most new work today is done using a variant of this type of model by macroeconomists of all political stripes and schools.
Now along comes two statisticians, Daniel J. McDonald and the acerbic Cosma Rohilla Shalizi. McDonald and Shalizi subject the now standard Smet-Wouters DSGE model to some very basic statistical tests. First, they simulate the model and then ask how well can the model predict its own simulation? That is, when we know the true model of the economy how well can the DSGE discover the true parameters? [The authors suggest such tests haven’t been done before but that doesn’t seem correct, e.g. Table 1 here. Updated, AT] Not well at all.
If we take our estimated model and simulate several centuries of data from it, all in the stationary regime, and then re-estimate the model from the simulation, the results are disturbing. Forecasting error remains dismal and shrinks very slowly with the size of the data. Much the same is true of parameter estimates, with the important exception that many of the parameter estimates seem to be stuck around values which differ from the ones used to generate the data. These ill-behaved parameters include not just shock variances and autocorrelations, but also the “deep” ones whose presence is supposed to distinguish a micro-founded DSGE from mere time-series analysis or reduced-form regressions. All this happens in simulations where the model specification is correct, where the parameters are constant, and where the estimation can make use of centuries of stationary data, far more than will ever be available for the actual macroeconomy.
Now that is bad enough but I suppose one might argue that this is telling us something important about the world. Maybe the model is fine, it’s just a sad fact that we can’t uncover the true parameters even when we know the true model. Maybe but it gets worse. Much worse.
McDonald and Shalizi then swap variables and feed the model wages as if it were output and consumption as if it were wages and so forth. Now this should surely distort the model completely and produce nonsense. Right?
If we randomly re-label the macroeconomic time series and feed them into the DSGE, the results are no more comforting. Much of the time we get a model which predicts the (permuted) data better than the model predicts the unpermuted data. Even if one disdains forecasting as end in itself, it is hard to see how this is at all compatible with a model capturing something — anything — essential about the structure of the economy. Perhaps even more disturbing, many of the parameters of the model are essentially unchanged under permutation, including “deep” parameters supposedly representing tastes, technologies and institutions.
Oh boy. Imagine if you were trying to predict the motion of the planets but you accidentally substituted the mass of Jupiter for Venus and discovered that your model predicted better than the one fed the correct data. I have nothing against these models in principle and I will be interested in what the macroeconomists have to say, as this isn’t my field, but I can’t see any reason why this should happen in a good model. Embarrassing.
Addendum: Note that the statistical failure of the DSGE models does not imply that the reduced-form, toy models that say Paul Krugman favors are any better than DSGE in terms of “forecasting” or “predictions”–the two classes of models simply don’t compete on that level–but it does imply that the greater “rigor” of the DSGE models isn’t buying us anything and the rigor may be impeding understanding–rigor mortis as we used to say.
Addendum 2: Note that I said challenge. It goes without saying but I will say it anyway, the authors could have made mistakes. It should be easy to test these strategies in other DSGE models.
Tuesday assorted links
1. Bob Solow making fun of Kydland and Prescott, via Dr. Ruth.
3. Craig Palsson, economics, and going viral on YouTube and TikTok.
4. AI and machine learning in economics.
5. Was the “Russian flu” in the late 19th century a rogue coronavirus? (NYT)
6. Twitter AMA with Marc Andreessen.
7. New Cecil Taylor jazz (NYT).
*Modern Paraguay: South America’s Best Kept Secret*
There should be more books serving as introductions to individual countries, and this one, written by Tomás Mandl, is a fine entry in the genre.
…Paraguay was South America’s first country to get electricity, railroads, and an iron foundry.
The Triple Alliance War of 1864-1870:
Although available data and sources remain contested, estimates put the figure at 25 percent of the Paraguayan population killed on the lower end, and upwards of 60 percent on the higher end…
For purposes of contrast, Poland during WWII saw “only” about 20 percent of the population killed.
Under Stroessner, the torture centers were neither secret nor undercover. And:
The clear pattern post-Stroessner is one of mild support for democracy: While in 2017 more Paraguayans agreed with the claim “democracy is preferable” than in 1995 (55 percent versus 52 percent, respectively), the average for the period was 46 percent….When Latinobarómetro asked Paraguayans to assess their country’s political regime on a range where 1 is “not democratic” and 10 is “fully democratic,” they have responded “5” consistently in almost every year of the twenty-first century.
With mandatory voting the average turnout rate is about 66 percent in recent times. And:
Notably, the largest center for Paraguayan studies is located in Argentina.
I enjoyed this sentence:
Unfamiliarity with Paraguay is not new.
Paraguay has very low FDI even by Latin American standards, it is typically rated as the most corrupt country on the continent, and a common saying is “¿Con factura o sin factura?”
Highly recommended, you can pre-order here, and yes the author does speak Guarani and he does also know the Solow growth model and why Singapore is interesting.
A Calculation of the Social Returns to Innovation
Benjamin Jones and Larry Summers have an excellent new paper calculating the returns to social innovation.
This paper estimates the social returns to investments in innovation. The disparate spillovers associated with innovation, including imitation, business stealing, and intertemporal spillovers, have made calculations of the social returns difficult. Here we provide an economy-wide calculation that nets out the many spillover margins. We further assess the role of capital investment, diffusion delays, learning-by-doing, productivity mismeasurement, health outcomes, and international spillovers in assessing the average social returns. Overall, our estimates suggest that the social returns are very large. Even under conservative assumptions, innovation efforts produce social benefits that are many multiples of the investment costs.
What was interesting to me is that their methods of calculation are obvious, almost trivial. It can take very clever people to see the obvious. Essentially what they do is take the Solow model seriously. The Solow model says that in equilibrium growth in output per worker comes from productivity growth. Suppose then that productivity growth comes entirely from innovation investment then this leads to a simple expression:
Where g is the growth rate of output per worker (say 1.8% per year), r is the discount rate (say 5%), and x/y is the ratio of innovation investment, x, to GDP, y, (say 2.7%). Plugging the associated numbers in we get a benefit to cost ratio of (.018/.05)/.027=13.3.
To see where the expression comes from suppose we are investing zero in innovation and thus not growing at all. Now imagine we invest in innovation for one year. That one year investment improves economy wide productivity by g% forever (e.g. we learn to rotate our crops). The value of that increase, in proportion to the economy, is thus g/r and the cost is x/y.
Jones and Summers then modify this simply relation to take into account other factors, some of which you have undoubtedly already thought of. Suppose, for example, that innovation must be embodied in capital, a new design for a nuclear power plant, for example, can’t be applied to old nuclear power plants but most be embodied in a new plant which also requires a lot of investment in cement and electronics. Net domestic investment is about 4% of GDP so if all of this is necessary to take advantage of innovation investment (2.7% of gdp), we should increase “required” to 6.7% of GDP which is equivalent to multiplying the above calculation by 0.4 (2/7/6.7). Doing so reduces the benefit to cost ratio to 5.3 which means we still get a very large internal rate of return of 27% per year.
Other factors raise the benefit to cost ratio. Health innovations, for example, don’t necessarily show up in GDP but are extremely valuable. Taking health innovation cost out of x means every other R&D investment must be having a bigger effect on GDP and so raises the ratio. Alternatively, including health innovations in benefits, a tricky calculation since longer life expectancy is valuable in itself and raises the value of GDP, increases the ratio even more. (See also Why are the Prices So Damn High? on this point). International spillovers also increase the value of US innovation spending.
Bottom line is, as Jones and Summers argue, “analyzing the average returns from a wide variety of perspectives suggests that the social returns [to innovation spending] are remarkably high.”
What I’ve been reading
Randy Shaw, Generation Priced Out: Who Gets to Live in the New Urban America. A YIMBY book, with good historical material on San Francisco, Los Angeles, and other locales involved in the struggle to build more.
Conor Daugherty, Golden Gates: Fighting for Housing in America. Coming out in February, this is a very good book about the YIMBY movement and its struggles, with a focus on contemporary California, written by a NYT correspondent.
Jennifer Delton, The Industrialists: How the National Association of Manufacturers Shaped American Capitalism. Why don’t more books fit this model: take one topic and explain it well?
Economists, Photographs by Mariana Cook, edited with an introduction by Robert M. Solow. Self-recommending. Interestingly, I recall an old University of Chicago calendar of economist photographs, still buried in my office somewhere, with pictures of Frank Hyneman Knight, Francis Ysidro Edgeworth, and others. At least in terms of personality types, as might be revealed through photographs, the older collection seems to me far more diverse. Or is the homogenization instead only in terms of photograph poses?
Michael E. O’Hanlon, The Senkaku Paradox: Risking Great Power War Over Small Stakes. A very useful practical book about what options a U.S. government would have — short of full war — to deal with international grabs by China or Russia. There should be thirty more books on this topic (#ProgressStudies).
Christopher Caldwell, The Age of Entitlement: America Since the Sixties. This is both a very old thesis, but these days quite new, namely the claim that 1965 and the Civil Rights movement created a “new constitution” for America, at variance with the old, and the two constitutions have been at war with each other ever since. It will be one of the influential books “on the Right” this year, I already linked to this Park MacDougald review of the book.
Robert H. Frank, Under the Influence: Putting Peer Pressure to Work. From the Princeton University Press catalog: “Psychologists have long understood that social environments profoundly shape our behavior, sometimes for the better, often for the worse. But social influence is a two-way street—our environments are themselves products of our behavior. Under the Influence explains how to unlock the latent power of social context. It reveals how our environments encourage smoking, bullying, tax cheating, sexual predation, problem drinking, and wasteful energy use. We are building bigger houses, driving heavier cars, and engaging in a host of other activities that threaten the planet—mainly because that’s what friends and neighbors do.”
Monday assorted links
Is the rate of scientific progress slowing down?
That is the title of my new paper with Ben Southwood, here is one segment from the introduction:
Our task is simple: we will consider whether the rate of scientific progress has slowed down, and more generally what we know about the rate of scientific progress, based on these literatures and other metrics we have been investigating. This investigation will take the form of a conceptual survey of the available data. We will consider which measures are out there, what they show, and how we should best interpret them, to attempt to create the most comprehensive and wide-ranging survey of metrics for the progress of science. In particular, we integrate a number of strands in the productivity growth literature, the “science of science” literature, and various historical literatures on the nature of human progress. In our view, however, a mere reporting of different metrics does not suffice to answer the cluster of questions surrounding scientific progress. It is also necessary to ask some difficult questions about what science means, what progress means, and how the literatures on economic productivity and “science on its own terms” might connect with each other.
Mostly we think scientific progress is indeed slowing down, and this is supported by a wide variety of metrics, surveyed in the paper. The gleam of optimism comes from this:
And to the extent that progress in science has not been slowing down, which is indeed the case under some of our metrics, that may give us new insight into where the strengths of modern and contemporary science truly lie. For instance, our analysis stresses the distinction between per capita progress and progress in the aggregate. As we will see later, a wide variety of “per capita” measures do indeed suggest that various metrics for growth, progress and productivity are slowing down. On the other side of that coin, a no less strong variety of metrics show that measures of total, aggregate progress are usually doing quite well. So the final answer to the progress question likely depends on how we weight per capita rates of progress vs. measures of total progress in the aggregate.
What do the data on productivity not tell us about scientific progress? By how much is the contribution of the internet undervalued? What can we learn from data on crop yields, life expectancy, and Moore’s Law? Might the social sciences count as an example of progress in the sciences not slowing down? Is the Solow model distinction between “once and for all changes” and “ongoing increases in the rate of innovation” sound? And much more.
Your comments on this paper would be very much welcome, either on MR or through email. I will be blogging some particular ideas from the paper over the next week or two.
Modern Principles of Economics
If you are teaching principles of economics next year do check out our textbook, Modern Principles of Economics.
Modern Principles means modern content and modern delivery. We cover material that many other textbooks ignore, such as how managers should choose between piece rates and tournaments and how firms can increase their profits using clever forms of price discrimination such as bundling and tying. In macroeconomics, we have created a simple yet powerful AD-AS model that combines insights from New Keynesian and Real Business Cycle models. We have also created the Super Simple Solow model which for the first time makes the Solow model of economic growth accessible to principles of economics students.
High-quality videos integrated directly into the textbook make Modern Principles a new kind of textbook, one born in the age of the internet. No other textbook has the quantity and quality of supplementary material available with Modern Principles. Whether you are looking to flip the classroom or just for the occasional video to grab the student’s attention before the lecture, you will find what you need in Modern Principles. The superb course management system also makes it easy to assign videos and grade questions.
Teachers can request a free examination copy here.
Here’s a video overview of Modern Principles of Economics:
Technology in Kubrick’s *2001: A Space Odyssey*
This post serves up some spoilers of detail, though no major spoilers of plot until the penultimate “you must go see it” paragraph. Upon a re-viewing of this movie, I found the following striking:
1. There is a Skype-like service for phone calls, but it never occurs to anyone that something like sending an email might be possible or even desirable. A lot of major and even apparently simple technological advances just aren’t that self-evident. The cameras in the movie also remained quite primitive and clunky, even by pre-smart phone standards. Maybe people expected a great stagnation in cameras back then.
2. At the time, Kubrick apparently thought it plausible that the audience would buy into common, widespread and indeed commercially viable space travel by 1992. The film was released in 1968.
3. Pan American flies people into outer space, and apparently used this new market to avoid total bankruptcy. Their stewardesses still have silly hats and costumes, and they act in a vaguely self-demeaning manner.
4. The film shows some signs of recognizing that Moore’s Law might happen. Hal for instance is advanced AI, but he is not huge in size. And the portrait of voice recognition technology is quite realistic.
5. Stars do not twinkle in outer space, however.
6. Hal 9000 would be less creepy with a female voice, and indeed Apple and Amazon figured that out some while ago. Note to my tech friends: do not program your personal assistant bots with a resentful, quivering, paranoid, passive-aggressive male voice.
7. The movie seems to suggest that chess-playing computers are a major achievement, when in fact this was mastered relatively easily, compared to many other AI problems. The movie shows this chess game, with Hal as Black. It is the kind of game you might expect a strong computer to play against a human, namely with a finish based on visually counterintuitive tactics.
8. It is a truly dystopic vision to think that Howard Johnson’s will be serving us food in space.
9. The first time I saw the movie, which I believe was in the mid-1970s, I was more stunned by seeing Americans talking to Russians “as if they were normal people” than by any of the technology.
Here is a good Wikipedia page on technologies in the movie. Now a few spoilers:
The movie, which I had not seen in many years, I found quite stunning. It took so many chances, and with so much self-confidence that the originality could be pulled off. Imagine opening a film with minutes of discordant Gyorgy Ligeti music, played against a dark screen, with no signal that this is even part of the movie. Then you see a long scene with apes, no dialogue to speak of, and no explanation of how this might fit into a commercially viable product. Finally the Solow residual is explained! There is not only no love story, the film arguably has no characters, Hal aside. Kubrick often expects ballet music to keep you interested, and various movements in space are stretched out to interminable length, yet almost always with striking aesthetic success. You could generously describe the ending as “underexplained.” Hardly anything happens in the movie, and yet at the same time it encapsulates the entire history of humanity with extra material on both sides, beginning and end, and a nod in the Hegelian direction.
Go see it on the large screen if you can — I can’t think of any film that is so much worse (or simply different) on TV as this one. It is one of the better movies ever made, and it dates from a time near Hollywood’s peak. It is sad that nearly two generations of Americans now do not know this creation as it was intended to be seen, and indeed must be seen. On 7 p.m. on a Saturday night, the theatre had no more than twenty people in attendance. When it comes to culture, salience usually matters more than you might think.
Our Textbook: Modern Principles of Economics
Writing about economics for a large audience at Marginal Revolution taught us to get to the point quickly, use vivid examples, and avoid unnecessary math and other jargon. We brought these skills to our textbook, Modern Principles of Economics. From the first sentence,
The prisoners were dying of scurvy, typhoid fever, and smallpox, but nothing was killing them more than bad incentives.
to the last, no other textbook teaches the economic way of thinking so well or so memorably.
Modern Principles means modern content and modern delivery. We cover material that many other textbooks ignore, such as how managers should choose between piece rates and tournaments and how firms can increase their profits using clever forms of price discrimination such as bundling and tying. In macroeconomics, we have created a simple yet powerful AD-AS model that combines insights from New Keynesian and Real Business Cycle models. We have also created the Super Simple Solow model which for the first time makes the Solow model of economic growth accessible to principles of economics students.
High-quality videos integrated directly into the textbook make Modern Principles a new kind of textbook, one born in the age of the internet. No other textbook has the quantity and quality of supplementary material available with Modern Principles. Whether you are looking to flip the classroom or just for the occasional video to grab the student’s attention before the lecture, you will find what you need in Modern Principles. The superb course management system also makes it easy to assign videos and grade questions.
Teachers can request a free examination copy.
Here are just two of our favorite videos available with Modern Principles. Enjoy!
From Achieve, the course management system, teachers can assign videos, homework, assessment, practice questions and more.
Achieve also contains an electronic version of Modern Principles with dynamic graphs. Dynamic graphs let the student interact with the textbook, giving students practice in shifting curves, for example, and seeing data presented in visually beautiful and informative ways. Here are just two examples.
Users of Modern Principles Love it!
“I can’t tell you how many people I have met who took economics in college, and who hated it. If only they had started with Cowen and Tabarrok. Modern Principles is one of the few books that will immerse students into the elegance and beauty of our science, and which will create a lifelong love of economics.”
-Lee E. Ohanian, Professor of Economics, UCLA and Senior Fellow, Hoover Institution, Stanford University.
“Cowen and Tabarrok’s Modern Principles and the accompanying videos make for an unbeatable combination for both students and instructors. The intuition is clear and the examples—both contemporary and interesting—draw students into the material. This text is a fantastic tool for showing students how economics impacts their daily lives in choices great and small. My students come to class with questions, eager to discuss in more detail the concepts covered in the videos and text.”
-Abigail Hall, Department of Economics,University of Tampa.
“I have tried multiple textbooks over the last ten years. None of them engage my students as well as Modern Principles by Cowen and Tabarrok. The writing is fresh and lively. The videos are clear and entertaining. It is a book that attracts students who will never take another economics course and excites economics majors.”
-Randy T Simmons, Professor of Political Economy, Utah State University.
Contact your rep to get more information.
Modern Principles, 4th ed!
Tyler and I are thrilled to announce the release of the 4th edition of our principles of economics textbook, Modern Principles. In the new edition we have fully integrated the microeconomics and macroeconomics videos that we have been producing for MRUniversity. No other textbook has anything like this wealth of supplementary material–putting it all together makes Modern Principles a new kind of textbook. We have also added a lot of new questions, Ask FRED questions, that use data from the FRED database, more material on health and economic welfare, more material on financial crises and fires sales and much more.
No other textbook has our super simple Solow model which for the first time makes the Solow model accessible to principles students. Modern Principles also has a balanced treatment of Keynesian and Real Business Cycle models, lots of material on modern topics like price discrimination including bundling and tying, a chapter on managing incentives (piece rates, salaries, tournaments) that’s great for MBA students and of course the best guide to understanding the marvels of the price system.
Check out the video!
GDP, GNP, and foreign investment
A few of you have written me to ask what I think of Paul Krugman’s recent posts on tax reform and evaluating it by gnp rather than gdp, the latter being an emphasis in the GOP literature. Paul notes correct that a lower corporate rate will attract foreign investment, and the returns to that investment, by definition, will not accrue to American citizens. So far, so good.
Paul reproduces the following graph for the Czech Republic, ratio of gnp to gdp:
If the GOP literature focuses on gdp, it is fine enough to criticize it on that basis. What worries me, however, is that the corrective doesn’t go nearly far enough. Gnp isn’t the right standard either, nor is gnp/gdp, rather it is welfare, either nationally or globally.
From that gdp/gdp ratio graph, you might come away with a grim view of life in the Czech Republic, but consider this cheerier picture of consumption, which nearly triples over a twenty year period:
Pretty awesome. And under the standard story of the Czech economy, investment from abroad, most of all from Germany, has helped drive those gains. Germany invested more, that boosted wages, improved the local political economy, and transferred some technology and entrepreneurial skills. It is standard international economics, or for that matter Solow model, that capital-rich, lower-return economies should invest in their poorer peripheries (which is not to say it always works out that way).
It’s entirely fair to note that Czech household debt to gdp has risen to about thirty percent. Still, in the U.S. it is about eighty percent, so the Czechs are not in dire straits just yet. Private debt to gdp seems to run about 136 percent, compared to about 200 percent in the U.S.
Of course, this still could end up as a bad deal for the Czechs. They might waste their foreign investment, the accompanying wage gains, the associated external benefits, and end up having to snap back their consumption and see their whole country owned by Germany, China, and others. But that’s not the baseline case. The default assumption is that these are gains from trade like other such gains, in this case gains from trading with foreigners who wish to invest. They are not lesser gains or gains to somehow be subtracted from the overall calculus.
Here is a useful point of contrast. Let’s say I advocated high taxes on foreign trade, on the grounds that “half of the gains from those trades are shared with foreigners,” and therefore we ought to, post-tariff, trade more with fellow citizens, so that only Americans get those gains. We all know why that argument generally is wrong, noting there are some second best cases where tariffs can improve welfare. It’s still wrong when the trades involve foreign investments.
So it is misleading to induce people to mentally downgrade foreign investment as a source of welfare gains. I get that Krugman doesn’t quite say that, but that is the impression his discussion and diagram produces on the unwary. Technically, he might only be criticizing the Republicans internally, using their own gdp standard. The actual produced impression is to cause people to doubt that a lot of foreign capital inflow fully counts as a gain from trade.
America of course is in a quite different position than is the Czech Republic. But the gains from foreign investment into the United States also ought not to be downgraded, either explicitly or by implied presumption.