May 29, 2011
Tyler Cowen: The Great Stagnation
Businessweek has a very flattering article on Tyler Cowen and how his book The Great Stagnation has prompted a shift in the debate about poor economic growth. Having read the book I strongly recommend reading it.
The Great Stagnation runs through three centuries' worth of what Cowen calls the "low-hanging fruit" of economic growth: free land, technological breakthroughs, and smart kids waiting to be educated. For developed economies, he argues, none of these remains to be plucked.
Even more important than land, in my view, is energy. As Peter Tertzakian pointed out in his book A Thousand Barrels a Second, previous energy transitions of the last couple of centuries were to energy sources that were cheaper and/or more convenient. The transition to petroleum combined both cost and convenience advantages. For example, the transition from coal to oil for heating eliminated the need to load coal into the furnace and to clean out ash. Oil enabled a huge growth in productivity of farm equipment, construction equipment, and in the transportation of humans and goods. The orders of magnitude increases in energy usage made possible our high living standards far more than did widespread college education. Many of the technological innovations were basically to use petroleum in more ways (e.g. to power jet engines). Our problem: Supplies of our most convenient fuel are looking pretty limited. What passes for good news on oil supplies (e.g. hydraulic fracturing of tight oil shale) costs more and projected flow rates aren't high enough to enable a return to the years of substantial annual increases in oil production.
"We thought we were richer than we were."
Yet America, Europe, and Japan have built political and social institutions on the assumption of endless growth. Cowen summarizes the financial crisis in eight words: "We thought we were richer than we were."
It's not that he disagrees with any of the better-known explanations for the crisis—easy credit, flawed ratings—it's that he sees a more fundamental problem, one that can't be fixed with regulation, bailouts, or tax cuts. Cowen thinks that now that America has used up the frontier, educated all of the farm kids, and built a couple of cars for every family, we might be done growing for awhile.
A slowing or halting of growth is highly problematic because so much has been set up based on the assumption of continued economic growth at a rate close the post-WWII average. Debt instruments have been issued and entitlements promised based on assumptions about long term growth that are not realistic. Pension funds project their funding needs based on rates of return from their stock market investments that are too optimistic. Even worse, we could be looking at something even more painful than stagnation: a receding tide of prosperity.
One can see signs of declining returns on investment in many areas. For example, doubling Pentagon modernization budgets delivered far less than was expected.
For almost a decade, the Defense Department saw its budgets boom but didn’t make the kind of technological strides that seemed possible.
“Since 9/11, a near doubling of the Pentagon’s modernization accounts — more than $700 billion over 10 years in new spending on procurement, research and development — has resulted in relatively modest gains in actual military capability,” Defense Secretary Robert M. Gates said in an address last week.
The article goes on to discuss what is wrong with US Army procurement methods. But I see something deeper at work to cause 10 years of US DOD spending at double the previous rate to yield only modest gains. The low hanging fruit are gone. The US military do not know how to get large gains from smaller scale projects. Therefore they try to do much larger scale engineering projects. Then many of those fail.
Computers are widely seen as huge enablers of progress and I agree they've delivered some big benefits. But what impresses me about the huge amounts of data getting sucked into computers and processed is that it has not (at least so far) enabled such better ways to manipulate matter in ways that can keep energy and other forms of commodities cheap. Investor Jeremy Grantham thinks we've going thru a transition into a much more resource-limited economic era. The industrialization of Asia combined with declining concentrations of minerals extracted from ores and other resource extraction limitations have caused a big turn in the direction of commodity prices starting in 2002 off of the long term trend of the last 100 years (see figure 2 there and my commentary here).
Some economists looked at the economic benefits of personal computers and assigned a small value of them. Compare those benefits to our far larger gains from coal- and oil-driven industrialization. I see computers as reducing (but far from eliminating) the pain of rising commodities costs by, for example, making engines run more efficiently and by reducing the need for physical travel. But computers so far haven't had an impact as great as the combination of oil drilling and mass-produced cars.
Okay, am I all doom and gloom? Nope. In spite of depleted fisheries, Peak Oi, declining ore concentrations, pollution, the potential for massive glacier melts and droughts, aquifer depletion, top soil loss, massive species extinction, human overpopulation (1970s environmentalists were right), and other problems too numerous to list I still hear Eric Idle singing. Potential very big benefits can be seen on the horizon from biotechnology. Very notably, we'll get new organs grown to replace old failing organs. But when will we get replacement organs, stem cell therapies, gene therapies, and great cures for cancer? So far all that's progressing much more slowly than the incredibly dramatic orders of magnitude declines in DNA sequencing costs.
Sequencing costs are going down at rates much faster than Moore's Law for computers. That can happen with really small stuff such as strands of DNA. But cells are much bigger and more complex things and organs are far more complex still. What I want to know: when will the advances in manipulating the small scale come fully to bear on stem cells, gene therapies and other key areas of biotechnology?
This analysis complete ignores strong AI. Once computers become auto-didactic, innovation will accelerate beyond society's ability to absorb it before it is outmoded by the next hour's technology. What then? Who would produce and sell products that are obsolete before they are sold?
Who would start a company when strong AI-driven personal fabs invent and produce every conceivable gadget a consumer needs, in the basement?
Nanotech, biotech, and strong AI will obviate business, industry, education, etc.
The benefit of computers has been to keep us from collapsing under the weight of our increasingly complex economy and society. Computers pretty much enable everything to work more smoothly, and have permitted us to grow to the size we are at right now. Now strong AI is almost within reach, probably 30 years, tops. If we can avoid total societal collapse for just 30 more years, strong AI will change everything. Anything is possible once we hit that mark.
There's this minor problem where we don't know how to build strong AI. Yes, you can sketch out a sort of hand-waving path where we manage to get there either via accelerated evolution of our existing AI or else a kind of incremental increasingly AI-assisted design process, but both of these are speculative. There is no guarantee that we will manage strong AI even if things don't fall apart.
I am tempted to read Tyler's book because I disagree rather strongly with his thesis (but of course I can't argue specifics without reading the book). I attribute the slowdown we see to demographics, increased regulation, and peak oil. Lack of free land? Really? First of all land has become an increasingly less important factor of production; second, we still have lots of unused land in the US. Maybe you can make a case for all the *really good* land being taken. As for smart kids waiting to be educated, well; yes, we don't have so many of those any longer, but that's down to demographics. I will grant the technological breakthroughs part, but it seems kind self-evident that developing new technologies gets harder as you go along.
We're seeing massive innovation in nanotechnology, none of which would be possible without computers. It will take time, yes, but there's going to be a tidal wave of new materials, dramatic changes to existing materials, etc. I think we're in a lull right now, but I don't think it will last more than a couple more years.
Strong AI is exciting but also unpredictable and scary. I don't want to be betting the future of humanity on that horse.
Raising human capacity through genetic technologies has more limited potential but it is also a safer bet. I suspect a modest programme would even be achievable now, if only somebody had the right combination of commitment, resources and know-how. Obviously, it would take many years to pay dividends.
Either way, time is running out. Without a breakthrough, the coming decades are going to see a massive contraction in population with all the horrors that entails.
Strong AI will not turn NP Complete problems into solvable problems.
To put it another way: Why do you assume more computation will yield solutions? We've had doublings of computational power every 18 to 24 months for decades and yet Western economies have slowed their growth and now have a hard time growing.
What is AI supposed to do? Near as I can tell: Write software for us. But how big are the search spaces for solutions?
I think computers have enabled the human race to go further into ecological overshoot.
Tyler also argues that fundamental innovations aren't coming as quickly. I think that's true. Theodore Maiman built the first laser on May 16, 1960. The first transistor was created December 16, 1947 by William Shockley, John Bardeen, and Walter Brattain. Name me a discovery of equal importance that happened after 1960.
Free land becoming less important: If that was true then the price trajectory of farm land would have been down over the last 10 years. Not the case.
Certainly much has been promised from nanotechnology. Nanotechnology, AI, and fusion energy are the big future promises.
One article on the Eagle Ford gets to the key point: "Just five years ago the oil fields were considered useless."
And there are 20 more oil fields waiting in the US. And probably hundreds more in the rest of the world that were once considered useless for oil.
This a new oil boom. A new Kuwait just discovered. Dozens more waiting to be rediscovered.
And that ignores the oceans of shale gas being discovered the same way.
This is what makes an interesting debate.
I suspect that the lack-of-energy card is overplayed.
If you look at the fastest growing large economy on the planet (China), they're effectively working off of a Victorian era transportation base: i.e. coal powered and relying on mass transit and ocean freight.
Why is it we think here in North America that we need an automobile based transportation system to grow?
I don't buy it.
In terms of resource restrictions: one point that's been missed is this; sure we've used up a sizeable portion of the easy to get to resources. Those resources have been in more temperate regions. Canada's far north and likewise the Northern Regions of Russia are where it's at. Historically those places have been inhospitable to humankind and as a result have been under explored. So there's still probably plenty left there.
Additionally, the growth we're likely to see going forward isn't simply use 2X the resources to get 2X the productivity. Instead it's likely to be a pinch of this, a pinch of that, all combined together.
Likewise: what do you need for biotech? Grow cells in the lab. They grow themselves as long as they are on a hospitable substrate, have oxygen and food.
Likewise demographics: why is it we think that the projected 9 billion is going to be much much worse than the current 7 billion? Most of the world already lives with way less than we do. But if you try visiting those places, the quality of life is more or less the same. The problems with "lower quality of life" countries is usually to do with lack of property rights and weak or oppressive governments leading to distorted markets rather than any supposed lack of resources.
An additional point about technological progress:
We can lose a hell of a lot of fat before we start digging into the meat and in the meantime, technology will continue to improve because the rich will continue to fund those technologies. The rich aren't going to disappear tomorrow.
Show me how the tight oil shales (not the kerogen shales which have had EROEI) are going to scale up to, say, 10 million barrels per day in the US and I'll agree happy days are here again. Is Eagle Ford going to scale beyond a half million barrels per day? Will Bakken go beyond 2 million?
Eagle Ford might have Oil equivalent to almost one third of current world annual oil production. So we need 3 Eagle Fords per year to maintain current world production rates.
"The extent of the oil play is quite large, more than 6 million acres of potential development," said Manuj Nikhanj, vice president of the Ross Smith Energy Group in Calgary, Alberta.
"You could eventually see 20,000 to 30,000 wells drilled in the play. You could have more than 10 billion barrels of oil through time. And the oil economics just keep getting better, so companies want to expand in this region
If Eagle Ford's production hits 420,000 barrel/day by 2015 then that would be about 2% of current US consumption. At $60 per barrel production cost we can afford this oil. Though it certainly demonstrates the extent to which the days of cheap oil are in our rear view mirrors.
The problem is incremental cost is going up. Add 2 billion more people and increase the buying power of a couple of billion others and the incremental demand increase runs up against declining return on capital outlay. Gotta use more marginal land, gotta use more expensive water sources, gotta extract lower quality ores.
Lack of energy card: look at the price of oil. That's the market speaking to us.
Most people do not see high rise apartment living as an appealing prospect. They see it as a decline in living standards. Also, most people see vacation road trips and air flights as how they want to travel.
China's Victorian era economy: A friend on a business trip had to pull her car over and throw up because the air made her sick. It makes the worst pollution LA ever saw as mild in comparison. China's now making more cars than the US per year and rising. They want to cruise too.
Hard to get places: Been happening for years. Mines in US, Australia, and other more easily accessed places are too depleted. The mining companies operate in Africa, Canada's north (hence the Ice Road Truckers), and other extreme places.
7 billion, 9 billion and the most of the world living with way less than we do: You are asking me a question and you are answering it. Those billions want more. We are going to get less as a result. We already are getting less as a result. We are using less oil than we were in 2006 and the Chinese and Indians are using more. it is a zero sum game. Add in 2 billion more and we will get even less.
To feed the industrializing Asians Western-style agriculture is going to get installed in areas of Africa that are now wild (get your trip in soon to see vanishing species). The incremental improvements in agricultural productivity are getting smaller in Western countries. Depleting soils and depleting aquifers are going to make this harder still.
Quality of life the same: I know people from India who do not want to live there again. Why? Too many people. I know guys who have gone over their to work to do meetings who see quality of life there as much worse. The roads are just jammed. People walk across 6-8 lane highways as the traffic comes to a stop. It is people at a density that Westerners can hardly imagine. I know a guy from Bombay who says "close the borders". Now that he's here he doesn't want any more people here. He's experienced extremely high density and wants none of it.
Fracking shale for oil is in its infancy. Its cost will fall just like the cost of fracking gas. Gas prices are down by more than 50% in the last few years.
Randall, a few months ago (weeks or days) we had people on this site denying that shale gas was important.
The North Slope is a coming shale oilplay, as is Haynesville etc etc
You see to forget than any industry that can switch to NG from oil will ... putting downward pressure on oil prices.
The problem is not that the NG and oil aren't there, its that the infrastructure to move it to market isn't there yet.
But that will great for jobs and steel mills making all that pipe!
All I needed to see was this.."David Brooks has called it "the most debated nonfiction book so far this year" and leaned on it for a column in the Times." The land of the insulated idiots gushing over something everyone else knew a long time ago. The farm boys educated went to engineering school and came out in engineering, their kids went into engineering for a semester and ended up in political science, now they come out with leisure management and a huge debt. The idea the west is a self-inflicted disaster is nothing new, Perot used it 20 years ago, and again as usual nothing will be done about it by washington.
Did the author discuss the stupid solutions coming from academia rather than just covering the obvious? The Fed has decided that wealth creation in the US now must come from an endless string of asset bubbles, destroying wealth all along the way, couple that with the present economic genius of building sidewalks during the day and taking them out at night, repeating until ultimately the world is just one big happy classless society, I bet another populist Perot pops up any day now, Trump was the precursor.
I first raised the issue on this blog in October 2009 on whether shale gas is a game changer. Since I see the answer as critically important I've read a lot on it since then. Still do not know for sure. But the longer natural gas prices stay low the more I'm inclined to believe.
I do not expect a big shift to NG to put much pressure on oil prices. Maybe oil will be $140 when it would have otherwise been $160. But still not cheap in any case.
Industries switching to NG: The critical question in my mind is what's the long term cost going to be for NGVs powered by CNG? The answer will help determine how much migration we see from outer suburbs toward denser areas.
A new term for the discussion: Available Net Exports (ANE), courtesy of Jeffrey Brown and Samuel Foucher. Take initial oil production, subtract off oil consumed by oil producers, then subtract off oil consumed by China and India. What's left is ANE. As you can see at that link the number is dropping.
What restrains economic growth now is politics. Just think how much better we could be if we removed the bag limit on lawyers.
Yes, essentially: The advances in productivity are all being eaten up for now in maintaining an ever higher predator/prey ratio between the political and productive classes. Things won't improve drastically until something happens to change that.
'if that was true then the price trajectory of farm land would have been down over the last 10 years. Not the case.'
Ten years isn't the timescale Tyler is talking about. The recent boom in farmland prices is driven by low interest rates and worldwide shortages of commodities. But anyway, it is true that cheap farmland would help; I just don't think it would help very much. Anyway if you wanted to talk about reasons why we are *about* to enter a period of slow growth, rather than why there has been slower growth since 1970, then I would agree that arable land and clean water are a factor.
The laser and transistor are interesting examples... it took quite a long time for either one to make a huge impact on the world (I would estimate roughly thirty years for each). Point being that there may be other things we've recently invented that we don't yet appreciate the magnitude of.
But what drives the worldwide shortage of commodities?
- The remaining oil costs more to extract.
- Ore grades are going down as they have for decades.
- Aquifers are depleting.
- Top soil is degrading.
- Fisheries are declining.
- Asian industrialization is creating a much larger base of demand.
Fundamentals far longer lasting than monetary policy are at work.
Tyler makes the case that innovations nowadays often benefit only a limited segment of the population.
As a historian, I certainly find wanting the idea that in the past new fuels were both cheaper and more convenient. On the contrary, new fuels could be crazily expensive provided they were convenient.
In XIXth cy, Britain imported at fabulous costs grease produced in tiny amount by South African sheep to burn in light houses: for a given mass, it lit better than coal. Oil (once used as ointment and medecine) was far more expensive than coal, but it drove engines both more compact and easy to use (anybody who has used a steam engine will appreciate this). Uranium was enormously more expensive than oil, but promised cheaper electrical power. Conveniency is what makes a fuel accepted, not cost.
I simplified somewhat. But look at your light house example. The amounts imported were small and for a purpose where the benefit was very large. The shift to a car-based culture required something far cheaper.
Oil was cheaper than coal for trains (and ships for that matter) due to other costs avoided such as refueling stations, labor to feed furnaces, and labor to clean out furnaces. Compare that to electric cars which require more frequent recharges and those recharges take much more time than putting gasoline in a car.
The shift from whale oil to petroleum was due to relative costs.
We now face an energy transition to sources that are less convenient and which have higher overall costs. Car batteries and Compressed Natural Gas (CNG) tanks cost more and the ranges are less.
QE2 was designed to inflate commodity prices and the stock market, an obvious clue to it's perverse effect on things in particular is gasoline prices are now as high at $100/bbl as they were when oil was $150 in early 08, just one more reason peak amazement at what the little agitator and his accomplices at goldman are getting away with has no end, US gasoline stocks just came off an almost record high, the end of january '11 had levels the highest in 20 years.
CNG is cheaper, and while ranges are lower, fillups are reasonably quick.
And there is lots of NG.
However ... with so much oil coming because of fracking I would be hard pressed to push the average car owner to switching early -- buying an NG car when the current one is done would be more prudent. Big trucks and fleet vehicles are the sweet spot for NG now, and in a few years as the infrastructure matures, then NG cars will be even a better choice than today.
A transition away from oil will reduce growth rates at least a little if for no other reason than it diverts scientific and engineering time away from new problems (like biotech) back to old problems we thought were fixed.
I've seen a claim that 50% of all engineers in the US work directly or indirectly on military applications. Much software and hardware may have non-military applications but can't exported for military reasons. What's the effect of the military drain on the US??
"50% of all engineers in the US work directly or indirectly on military applications."
Really ... all of these occupations are working for the military? http://www.bls.gov/oco/ocos027.htm
I heard unicorns are real too.
As Abraham Lincoln once said: Don't believe everything you read on the internet.
Civil engineers 278,400
Mechanical engineers 238,700
Industrial engineers 214,800
Electrical engineers 157,800
Electronics engineers, except computer 143,700
Computer hardware engineers 74,700
Aerospace engineers 71,600
Environmental engineers 54,300
Chemical engineers 31,700
Health and safety engineers, except mining safety engineers and inspectors 25,700
Materials engineers 24,400
Petroleum engineers 21,900
Nuclear engineers 16,900
Biomedical engineers 16,000
Marine engineers and naval architects 8,500
Mining and geological engineers, including mining safety engineers 7,100
Agricultural engineers 2,700
Engineers, all other 183,200
Simply presenting numbers without context doesn't tell us much. The US military is a $900B operation, and it is heavily concentrated in high tech purchasing: airplanes, weaponry, unique things like tank drive trains etc, etc. Nuclear engineering and the internet started life as military applications - so did the interstate highway system. Those have all developed a "civilian" application, but they took a while to do so, and many technologies never do. How much engineering talent is developing weapons systems right now? A whole range of related things, like
Boeing, for instance, is a very important military contractor - so much so that the EU provides massive subsidies for Airbus to level the playing field.
Randall - how much software do you see that is classified as not legal to export?
The military component of design engineering is surely much less than 50%. Only part of the DOD's budget is procurement ($140 billion? something near that) and only a part of procurement is engineering.
In Bruce's list we can easily see that some whole categories have nothing to do with the military (mining, petroleum, ag, environmental - except for cleaning up weapons factories).
Look at Boeing: What percentage is military? My vague memory is it used to be about half and half. I'm guessing civilian is bigger for them now.
Software not legal to export: I never come across that in the industries I've worked for in recent years. Sure, there are some countries that are not legal to sell to. But they aren't legal to sell much of anything to.
Nick G: "Simply presenting numbers without context doesn't tell us much."
I think there was more context in my list than your 50% claim.
Nick G: "How much engineering talent is developing weapons systems right now?"
Interesting question. Way less than during WWII or the Cold War would be my answer. Why don't you do some research?
I don't have the time to research it - I thought it was worth asking the question in case someone happened to have some concrete information.
It's possible to have a real debate while being respectful of others. I know it's tempting, but being disrespectful doesn't win arguments - it weakens them, really, and makes the whole thing unpleasant.
Making stuff up and then asking other people to do the research is pretty disrespectful.
I didn't make anything up - I asked a reasonable question based on a claim I'd seen.
Actually, you and Randall both gave useful information, even if it was far from definitive. It's seems reasonable to think that the 50% figure is either too high, or it referred to a sub-category of technical work/engineering.
The more important point: military spending and R&D can weaken the economy. A number of analysts have said that:
Yale historian Kennedy surveys the ebb and flow of power among the major states of Europe from the 16th centurywhen Europe's preeminence first took shapethrough and beyond the present erawhen great power status is devolving again upon the extra-European states. Stressing the interrelationships among economic wealth, technological innovation, and the ability of states efficiently to tap their resources for prolonged military preparedness and warmaking, he notes that those states with the relatively greater ability to maintain a balance of military and economic strength assumed the lead.
The Rise and Fall of the Great Powers - Paul Kennedy
Just how much it does so is an interesting question.
Finally, I understand the temptation to take an aggressive tone - I've done it too... but please don't. It's no fun.
You know, heard on the internet you were an idiot Nick. Research seems to back me up.
Name me a discovery of equal importance that happened after 1960.
In a few years, I'm pretty sure it'll be agreed that it's graphene.
Will Bakken go beyond 2 million?
At 3.65 billion bbl technically recoverable, 2 mmbbl/d would exhaust it in 5 years. Given what's economically recoverable and the need to amortize infrastructure, 200 mbbl/d peak over ~30 years is more likely. That means we'd need 50 Bakkens to get 10 million bbl/d... and we have one.
Bruce calling someone else an idiot is pretty ironic, the Dunning-Kruger effect in action.
E-P has taught you well to be unnecessarily rude even when the other person wasn't rude to you.
Don't ever say Bruce is incapable of learning.
However, I will argue that Bruce's appearing from nowhere and e.g. insulting everyone who's not an AGW-denialist is extremely rude even if he didn't address anyone in particular (which he did, early and often).
All the smart kids are educated? I beg to differ. There are cheap, easy, and tremendous returns possible from ACTUALLY educating all of the underproductive youth who are nominally-educated-but-functionally illiterate.
And then drastically reforming our education system so we turn out peole with the intellectual toolkit needed for a modern society.
Comment-of-the-decade was made on this thread by commenter named 'th' who, on May 30th 2011 at 6:43 PM (see the comment above) predicted the rise of Donald Trump in American politics, going so far as to reference Trump by name. Lauds, th.. ---the Eschaton.
Thanks for pointing that out. Yes, th did predict a new populist and mentioned Trump. Pretty cool.