|Book Review: The Great Stagnation
||[Jan. 19th, 2013|10:31 pm]
All these plane trips lately have meant lots of books getting finished, and the latest is Tyler Cowen's The Great Stagnation. I wanted to like it because I like Cowen's blog and the book's thesis - that the world's technological and economic progress is slowing to a halt - appeals to the crotchety catastrophist in me. Instead, I thought the book was grating to read and didn't defend its main thesis well enough to be worthwhile.
Cowen believes that the era of spectacular economic and technological growth lasting from the Industrial Revolution to about 1970 is now over, and right now we're just sort of wringing the last couple of drops out a few nearly-empty low-hanging fruits from the past (this mixed metaphor is in honor of Cowen, whose book contains cringe-inducing overuse of the phrase "low-hanging fruit").
He compares the many spectacular advances of the pre-1970 world - steam engine, railroad, camera, combustion, telegraph, telephone, electricity, light bulb, radio, television, automobile, airplane, computer, jet engine, et cetera - to the relatively few advances of the post-1970 world, which he pretty much just lists as "the Internet". He blames this on a few things, including increased spending on unproductive activities, screwed up politics, and of course mining all the veins of low-hanging fruit to exhaustion.
The book is well worth thinking about, and even well worth reading despite what I considered the awkwardness of its prose. But in the end I think I reject the thesis. Let me explain why.
Kurzweil Meets Anti-Kurzweil
Here's a graph very similar to one in the book. It shows median US income over time. We can see that up until about 1970, it increases rapidly. Around 1970, it shifts into a holding pattern with occasional fluctuation but no general upward trend. So far, so good for the Great Stagnation thesis.
Here's the same graph, except it's added GDP per capita, a measure based on the mean rather than the median. Mean GDP per capita stays at exactly the same upward trend throughout. This graph was emphatically not in the book, and I am puzzled as to why. Tyler Cowen is a very smart economist and I'm sure he knows about this. I am very reluctant to contradict him about economic data which is something I know almost nothing about. But since I expect I will be corrected if I'm wrong, let me explain what I think the implication is.
America is growing at about the same rate it's always grown. Starting in the 1970s, the growth became concentrated in the hands of a few very rich people, who became very very rich. This lowered the growth in median income without affecting the mean.
If this is true, there's not a Great Stagnation, there's a Great Widening Of The Gap Between Rich And Poor (note to self: come up with better name for this before writing a book on it). And in fact we know there has been a Great Widening of the Gap Between Rich And Poor, and that it did start around 1970. This seems to provide an almost-fatal blow to the Great Stagnation theory; it's really hard to argue that growth is stagnant when the GDP growth chart sticks exactly to the trend line (this is even clearer here) without 1970 having even the slightest kink.
(Once again, this seems like an obvious enough objection - and one none of the other reviewers raised - that I expect I am missing something here.)
But if growth is stable, what about the argument from diminishing inventions? Cowen makes this argument in two parts. The first part relies on the research of Jonathan Huebner, who has been tracking various measures of innovation over time:
Okay, first things first, that is one really suspicious trend line on that first graph. Not quite corporate taxation Laffer Curve-level suspicious, but suspicious nonetheless (just to make it clear I'm not straw-manning here: "'We are approaching the dark ages point, when the rate of innovation is the same as it was during the Dark Ages,' Huebner says. 'We'll reach that in 2024.'"
Second of all, that "all the innovations since 1400" graph is very vulnerable to what you count as an innovation. Kurzweil tries almost the same exercise and uses it to prove the rate of technological advance is going up.
The second graph uses actual US patent data, which is both better and worse. It's better because number of patents is an objective measure that's hard to cherry-pick. It's worse because measuring science by patents seems a lot like measuring art by weight ("We produced sixty tons of oil paintings last year!"). Nevertheless, even on its own terms, this graph fails to support the book's hypothesis: it looks from the graph like the rate of patent discovery has been going down since 1930 - right smack in the middle of Cowen's Era of Low-Hanging Fruit In Every Garage - and going slightly up since 1970. But in general both eras seem to be following a line of which Kurzweil might well approve.
This is a chart not of patents per million people, but of patents in general. It tells a much clearer story, which is that patents have been going up at what looks like a very continuous (and fast - the graph is logarithmic in scale!) rate for as long as they've been measured with the exception of the Great Depression. Once again, the era from 1970 to the present looks exceptionally good. We can also find that the growth rate of science continues to be freakishly steady at 4.7%/year, although the methodology used for that calculation is even worse than measuring art by weight, more like measuring art by number of donkeys needed to haul it or something. But it only confirms what all our other statistics are showing.
If we can't prove the Great Stagnation with the math, what about our own eyes? Like Cowen, I have an easier time naming revolutionary inventions of the past than revolutionary inventions of the present. But this is to be expected: revolutionary inventions of the present do not seem like revolutionary inventions yet.
When William Shockley invented the transistor (with some help from a probable distant relative of mine) in 1947, the news headlines did not read "MOST IMPORTANT INVENTION OF TWENTIETH CENTURY DEVELOPED". If they existed at all, they probably said "You know those giant machines that the government sometimes uses to break codes? It looks like someday we might be able to build one of their component parts more efficiently. Who knows if one day our great-grandchildren will marvel that we were alive at the time they invented the polywell fusor or the photonic crystal, and we'll sagely nod our heads and say "Yeah, I think they mentioned something about that on Twitter once."
It may even be unusually difficult to notice anything being invented when you're in the midst of it. One day they say they've discovered quantum computers are possible, and you think "Yeah, an existence proof, big deal." A few years later they've successfully sustained a few qubits and you think "Yeah, still far away from a practical application." Then later they build a very clunky prototype quantum computer, and you think "Just a natural progression of the qubits they already had, and it's not even as good as a regular computer yet." And then a few years later you've got a commercial quantum computer on your desk that can calculate the square root of infinity in half a second, and you just think "Well, it's about time, considering how long the government and academics have had these." At no point do you think "Whoa, a quantum computer, this is so futuristic!"
And finally, in order to make his argument work at all, Cowen has to say "Let's ignore the Internet." Really he needs to ignore all electronics: the personal computer, the cell phone, the smart phone, the video camera, the mp3 player, voice-over-IP - but yes, the Internet is the most glaring counterexample. Cowen here notes that the Internet, for all its importance, hasn't had much direct effect on economic growth. Wikipedia has made life immeasurably better, but it hasn't pushed a lot of products, employed a lot of people, or consumed a lot of widgets in the same way a General Electric or a Ford Motor would. And to his credit, Cowen says that this is more a condemnation of the way economists measure value than of Wikipedia. But he doesn't take the next step and allow the existence of all this nifty online technology to count against his Great Stagnation hypothesis.
And to be fair, you would think if we were really technologically advancing, we would see these advances in fields other than computing. Maybe computing is just so fascinating right now that a disproportionate number of our smart scientists, rich venture capitalists, and social attention is on it right now and it's distracting us from doing other things? I dunno. But saying "Aside from computing, the 21st century has been pretty lackluster" is a lot like saying "Aside from industrialization, the 18th century was kind of unimpressive."
This Time It's (Not) Different
Another piece of strong evidence came in the form of very recent economic indicators - all of which are depressed because of the recent recession. Cowen (or possibly not Cowen and just other stagnationists; I don't remember) claims that this is the chickens finally coming home to roost: that rather than Yet Another Recession In The Business Cycle, this is a population drunk on low-hanging fruit finally being forced to come to terms with the new normal.
This is a convincing argument, but...
I have now lived almost thirty years. I have been in a few booms and a few recessions. Every time, I have heard people make very clever and convincing arguments why this time the boom will never end, or why this recession is structural and inescapable. Each time they are just as persuasive as the last. And each time, the boom or recession ends, much like all the others, and there is much gnashing and wailing of teeth from those who had bet otherwise. In fact, I have come to Notice A Pattern. And I have read books from people older and wiser than I, and they have confirmed that they have Noticed A Pattern too.
And some have suggested that before the boom or recession even begins, one should make an Outside View inspired resolution to just completely ignore all the claims that This Time It's Different. And this seemed wise, and so I did that.
This is not a foolproof strategy. It may be that Sometimes It Really Is Different. The Japanese financial crash of 1990 or so was different, and Japan never really recovered from that. Some people have said this crisis here is very similar and will do the same to the US. They make good points. But I'm deliberately ignoring them.
It's also kind of hard for Cowen to use this recession to support his thesis when we have a pretty good idea what caused it and stagnant technology wasn't it.
Throwing Money In A Giant Human-Welfare-Shaped Ditch
The strongest part of the book in my opinion was the chapter on how much money we're throwing away.
Most people here will be familiar with Robin Hanson's work on health care. Cowen re-iterates that work. Health care spending is growing almost ten percent a year and the percent of GDP devoted to health care has doubled since 1980. This might be acceptable if we were getting better health care for our money, but we very clearly are not: aside from changes in infant mortality rates, life expectancy has barely budged in the last few decades. Studies, especially the famous RAND study, show that people who spend more on health care don't have significantly better health outcomes than anyone else. A very large amount of health care - between a quarter and a third - is spent in the last year of a person's life, trying to hold off the inevitable.
Health care is now 16% of the US GDP and this number is only going up with no end in sight. It's diverting resources that could be going to infrastructure and new non-health technology and so a massive drain on everything. I would say that the waste was astronomical, except that as Feynman points out there are only a few billion galaxies in the universe and so once you reach trillions of dollars of waste it is more accurate to just say the waste is at levels-only-found-in-economic-waste-numbers.
What I hadn't heard before, and what makes a similar amount of sense, is the same argument with education. Since 1970, per-pupil spending in school in real money has doubled, with practically no effect on most measures of student scores. Schools that spend more per pupil don't get any better results once confounds for economic class have been adjusted out. We spend about 150% more money on education than Germany but aren't doing any better than they are.
I am not including a chart in this section because it confuses me. The conservatives put out charts like this one which purport to show massive increases in per pupil school spending, but graphs like this one show total spending as percent of GDP as quite stable. I'm not sure what to make of that since I would expect the inflation adjustment in the former to do about the same work as the GDP-percent-ing in the latter.
Still, a government constantly spending more and more money on education without seeing any better results seems to fit with the way I expect the world to work pretty well.
Cowen's arguments on health and education are important both because they're things more people should know about - in the sense of "we're spending huge amounts of money and not getting anything in return" - and also because they provide a way his Great Stagnation thesis could still be sort of correct even given our high GDP.
Overall I disagree pretty strongly with the rest of the book. The evidence from GDP and scientific output seems to pretty thoroughly disprove a stagnation - it would be possible to question it, but I can't think of on what grounds you would even want to.
I was also turned off a bit by the bloggish style of the book; I like Tyler Cowen's blog, but hearing a blog tone in a book was kind of weird. And the last chapter - where Professor Cowen tells us it's all going to be okay after all, and gives some helpful hints - it just seems totally incongruous. The assertion that everything will get better seems flimsy and unsupported, and the helpful hints, like "We should give more status to scientists", see like Vaguely Good Things but not really related to the thesis of the book.
I will happily praise the book's few forays into politics, as well as its excellent job not foraying into politics more than it did. It took a very even keel between liberalism and conservativism, even though it would be very easy to use its main thesis in praise of either side. And the few political points it did make - about the uselessness of health and education spending, and about how our politics is pretty much built on an foundation assumption of constant growth that can be siphoned off to fund programs to bribe voters - is well taken. Even if the constant growth hasn't stopped quite yet, that's still a scary foundation and one we need to be thinking about.
And I will not as happily, but still honestly, admit that even though I found the book's arguments weak, I feel strangely compelled to believe it.
I don't know what is so attractive about the Great Stagnation thesis. Maybe it's a sort of morality story, a feeling that our politicians, business leaders, and voters have screwed things up so bad that it would be only fair if we started stagnating; they've been beating the goose so hard that one almost wants to see them get their comeuppance when the golden eggs stop. That there's so much hypocrisy and irrationality that the ability of the economy to just keep churning on and on at its accustomed pace is almost proof that nobody cares about our sins and by extension our virtues. Maybe it's the hope that if something went horribly wrong, we could finally get about smashing the system and building a more elegant one, which is always attractive to think about whether or not it's merited.
Or maybe it's that you hear stories about the old days, when two bicycle mechanics outside any formal academia or Pentagon research laboratory could watch some birds, think really hard, and invent the airplane through nothing but pluck and Good Old-Fashioned American Ingenuity. And the modern world of establishment academics doing research for big corporations is so disappointing in comparison that it would be better for it to be a dead end than for it to be the shape of things to come.
There's another thing that bothers me too, which is the sheer damnable linearity of the economic laws. The growth of science as measured in papers/year; the growth of innovation as measured in patents; the growth of computation as measured by Moore's Law; the growth of the economy as measured in GDP. All so straight you could use them to hang up a picture. These lines don't care what we humans do, who we vote for, what laws we pass, what we invent, what geese that lay long-hanging fruit have or have not been killed. The Gods of the Straight Lines seem right up there with the Gods of the Copybook Headings in the category of things that tend to return no matter how hard you try to kick them away. They were saying back a few decades ago Moore's Law would stop because of basic physical restrictions on possible transistor size, and we worked around those, which means the Gods of the Straight Lines are more powerful than physics.
And the lines slant upward, which is very kind of them, but not all of them make sense, and none of them seem to be in our very direct control, and some of them are on log plots, and some of them, when drawn out in a Kurzweillian fashion, go to kind of scary places.
So no, I don't think it is going to be a Great Stagnation we are going to have to worry about.