Are we doomed to drown in stuff, or run out of the raw materials to make it? After the midwinter potlatch I was ready for some good fire and brimstone on this well-worn theme. George Monbiot is usually a reliable Savonarola, but I found his latest Christmas diatribe against growth and consumerism disappointing.
Every Friday is a Black Friday, every Christmas a more garish festival of destruction. Among the snow saunas, portable watermelon coolers and smart phones for dogs with which we are urged to fill our lives, my #extremecivilisation prize now goes to the PancakeBot: a 3-D batter printer that allows you to eat the Mona Lisa or the Taj Mahal or your dog’s bottom every morning. In practice, it will clog up your kitchen for a week until you decide you don’t have room for it. For junk like this we’re trashing the living planet, and our own prospects of survival. Everything must go.
Personally, I’d have gone with the Great Pacific Garbage Patch, now twice the size of France. [Update: you can of course follow her on Facebook.] The trouble with anecdotes, however stomach-churning, is that they don’t tell you anything about the trend. Are Monbiot’s ghastly examples typical, or the reflection of the fact that most middle-class people in rich countries already have all the material possessions they need and most of what they want? In that environment, finding affordable presents the recipients will actually like is getting harder and harder, before we finally stop the pointless exercise.
For the trend, we need models and numbers. I’ve already written about solid research by Thomas Wiedmann et al that Monbiot pointed me to, showing that:
1. The material intensity of world GDP has been going down.
2. It is still coupled to GDP, and there is no complete dematerialisation of growth.
So far so so-so. Wiedmann’s data stop in 2009, and he hasn’t updated yet. To fill the gap, Monbiot pointed to a new paper by Australian economist James Ward et al, purporting to show that decoupling of economic growth from material inputs is an illusion. IMHO this is question-begging hothouse orchid-growing; a wearisome takedown below the jump.
To get an idea of what’s been happening recently, I had a go myself with rustic methods. I constructed indices of world consumption for five significant materials (steel, aluminium, copper, cement, paper and board), normalised to 2005, before the financial crisis. Global CO2 emissions from fossil fuels track their production. I couldn’t find world data for construction sand and gravel, so I threw in data for the USA: it’s interesting because these are very cheap and widely available, so consumption cannot be significantly affected anywhere by price or supply constraints. Here’s the result. A pretty chart; spreadsheet with working and data sources here. [Update 5 April: following a suggestion by commenter Nick at John Quiggin’s blog, I’ve added gross shipping tonnage to the charts and spreadsheet.]
This clearly suggests that the partial decoupling established by Wiedmann has got stronger recently. The inflection seems to have happened around 2013.
This is pretty crude, but it tells a clear story. Raw materials only lagged GDP by a little up to five years ago. Since then – and without any global recession – they are running at about half of GDP growth, and only a little faster than population. There is just one outlier in my basket, aluminium, which is still replacing steel as lighter and more durable. We have not yet quite reached Peak Stuff. But the strong dematerialising forces that created the moderation are still at work. It is a reasonable hope that we will pass the peak in the coming decade.
Timeline of my posts on this topic, changing my mind twice (new information, you see, and no help from the stars of the economics profession):
Long comment on the James Ward paper below the jump.
The James Ward paper cited approvingly by Monbiot is mainly an exercise in oversimplified a priori modelling. There is also a detailed analysis on Australia, a small economy unusually dependent on mining. This may be good stuff, but it is no basis for inferences about the world. For that, it’s entirely the grand model.
To warm up, a few lesser snarks.
One: they lump energy inputs with material resources and land. This is unsound. If our future energy comes from fossil fuels, the greenhouse effect from growth will destroy civilisation in a century, long before it runs out of anything else. Contrariwise, if energy supplies are renewable, these are unlimited for practical purposes. Ultimately the material requirements of wind turbines, solar panels and batteries would impose an upper limit; but it’s so far above the current extravagant energy use of rich countries, already flatlining, that it’s surely not a worry at a reasonable planning horizon of a century.
Two: they have a hazy understanding of the decoupling hypothesis.
However, there are several limitations to the inference of decoupling from national or regional data. There are three distinct mechanisms by which the illusion of decoupling may be presented as a reality when in fact it is not actually taking place at all: 1) substitution of one resource for another; 2) the financialization of one or more components of GDP that involves increasing monetary flows without a concomitant rise in material and/or energy throughput, and 3) the exporting of environmental impact to another nation or region of the world (i.e. the separation of production and consumption).
Point 1 is baffling. What’s the metric of material input? Wiedmann, drastically, uses weight. This creates some anomalies, especially over natural gas, but it has the merits of clarity and observability. Is there any reason to think that massive substitution of lighter for heavier resources is taking place? There is gas, of course: but as I’ve argued above, we need to look at energy separately. Some products, like computers, are getting lighter, but that’s not what they mean. The call by Mike and me for a major shift from steel and concrete in construction to lighter timber (advert) has not yet been taken up at any scale.
Alternatively, you can weight material resources by scarcity. Substitution of common resources for rarer ones has been a regular feature of the world economy since the replacement of scarce copper and tin (for bronze) by abundant iron in the first millennium BC. A striking recent example is the complete replacement of gold and silver as means of payment by paper banknotes (dirt cheap and indefinitely sustainable) and dematerialised bank deposits, merely entries in electronic ledgers. We are not going to run out of iron, paper, or bytes. This substitution is an integral part of technical development and the dematerialisation story, not some statistical fiddle.
Point 2 is absurd. Granted that the growth of the financial sector has brought few if any benefits to the general welfare, and much of it is welfare-negative rent-seeking and zero-sum speculation. But the sales of say credit-default swaps are indubitably part of GDP, and they indubitably have very little material footprint. It’s decoupling at work, like it or not. The claim is that GDP can’t grow forever, not some other cuddlier indicator they have not defined.
Point 3 is perfectly true, as shown by Wiedmann. But his data show a relative decoupling in China, even after the offshoring effect. China’s economy is tertiarising fast, and both steel and coal production are down, even accounting for exports. China may or may not continue on this path, but it’s an empirical question Ward et al do not address.
Now for the main course.
They start OK.
Put simply, absolute decoupling from resource or pollutant j requires Tj [the input intensity in appropriate physical units of the resource per unit of GDP] to decrease by at least the same annual percentage as the economy is growing.
Fine. But why can’t it decrease forever at less than the growth rate forever of GDP, shrinking to an asymptote at zero? They go on (my emphases):
For non-substitutable resources such as land, water, raw materials and energy, we argue that whilst efficiency gains may be possible, there are minimum requirements for these resources that are ultimately governed by physical realities: for instance the photosynthetic limit to plant productivity and maximum trophic conversion efficiencies for animal production govern the minimum land required for agricultural output; physiological limits to crop water use efficiency govern minimum agricultural water use, and the upper limits to energy and material efficiencies govern minimum resource throughput required for economic production. Therefore a more appropriate formulation of Eq (4) is to allow Tj to decrease to an ultimate value, Tult ≥ 0, as follows:
where Tj,ult is the ultimate resource use intensity, and rj is the rate of exponential decline, for resource or pollutant j.
So as GDP expands, at some point it hits the minimum intensity and resource uses expands in lockstep with GDP, towards the inevitable resource Armageddon. QED!
Our model demonstrates that growth in GDP ultimately cannot plausibly be decoupled from growth in material and energy use, demonstrating categorically that GDP growth cannot be sustained indefinitely. It is therefore misleading to develop growth-oriented policy around the expectation that decoupling is possible.
The crucial mistake is the first, “intensity”. This entirely begs the question by assuming the desired result. The decoupling hypothesis indeed requires a continued exponential decline in material intensity of GDP. It is compatible with absolute floors in input of material resources in relation to physical quantities of use, for example per head of population, per car, per flight, as long as these themselves severally reach an absolute peak. (We confidently expect world population to peak later this century). The same holds for any material object you can think of, like a car or a phone. The decoupling thesis is that there is no limit to the shift in the composition of output away from such material goods. The limits-to-growth crowd are as fixated as Soviet central planners on the vision of production as stuff. Information and services are GDP too, and a rising share of it. The dematerialisation of growth hypothesis is coherent, and it’s an empirical question whether or not it’s true.
An example may help. Consider a typical example of a product in the tertiarised information economy, a professionally produced and marketed cinema film. Worldwide, the production cost of a 90-minute film ranges from the $50,000 of Nollywood soaps shot in garages in Lagos using Sony handycams to the $378 million of Pirates of the Caribbean: On Stranger Tides (2011).
These are extreme cases. To get a more typical range, take the filmmaking of Woody Allen and Pixar, both consistently successful. (Allen is not a nice man, but he is undoubtedly a highly competent professional). For three Allen films for which production budgets were released, we have an unadjusted average of $18m, and a range of $15m – $22m. For sixteen Pixar full-length animations, we have an unadjusted average of $128m,, and a range of $30m – $200m.
Pay attention next time to the roll of credits at the end. Even a cheap art house movie like one of Allen’s involves a hundred or so skilled workers, ranging from decently paid drivers to exceedingly well-paid leading actors. The screen credits are incomplete; the accountant may just make it in, but not the accountant’s secretary or the cleaners. Pixar’s credits are much longer, as the filmmaking requires a small army of animators. Contrast this huge amount of cooperative labour to the limited material footprint. Pixar works entirely indoors, in Dilbertian cubicle offices stuffed with hundreds of state-of-the art workstation computers. Allen shoots in conventional studios and on location, where a large team has to be flown in with much of its equipment. There is no reason to think the material or energy footprint of this activity is increasing over time in step with its commercial value.
In one respect, the footprint is decreasing quite markedly. Pixar animations are all digital. Allen switched from chemical stock to digital in 2015. Digital films are recorded on cheap and reusable storage media, ultimately hard drives. The film itself is pure information: trillions of noughts and ones, making up an immaterial object worth up to $200m. The cost per copy for distribution falls from ~$2000 to ~$50, with far better IP security. Digital films still presumably have to be transferred to chemical stock to match legacy projectors in cinemas, but digital projection is spreading fast and will ultimately make chemical stock obsolete. Nollywood never uses it anyway, the films are distributed on 50 cent video CDs.
The full decoupling hypothesis requires the deconstruction of GDP into its material and immaterial components, and makes the claim that the material components have low elasticity w.r.t income and can and will go flat. It appeals to a segmented and hierarchical view of human needs and desires; this isn’t outré, it’s in the stock Maslow pyramid of needs from 1943. Taken separately, wants are satiable. In rich countries, very few people have to worry about getting enough food, shelter, or warmth. The satiation of basic economic wants reveals others: nice food cooked and served by other people, comfort, entertainment, learning, art, status, space, privacy.
It’s a nice question whether the last three are satiable. Relative status is zero sum, so it for one cannot be. Space and privacy are limited in a world of 10 billion. It’s possible that beyond a usable maximum (represented say by a luxury flat in New York or London) the demand for space is just a form of competition for status, a positional good. These difficulties deserve investigation, but at present do not prove that in the aggregate material decoupling is infeasible.
I’m pretty confident, on the basis of the flatlining of material product in OECD countries and its converging slowdown in China, that dematerialised growth is feasible for a good while.
For the long term, I dunno. The vision of the future as a giant Versailles with static material inputs, a never-ending expansion in the elaboration of cultural consumption, and a bitter Darwinian conflict for seats near the throne is unattractive and does not look stable. We don’t know if the full communist alternative of “help yourself to everything” is feasible, as in the late Iain Bank’s far-future galactic Culture, with the full menu of supertechnologies – sentient AI, FTL drives, matter transmutation, you name it. Communism could surely work for the contents of Chinese hardware stores. Who looks at the prices? If you need a screwdriver, you get one. There is also scope for expanding the quite large area of life in which gift exchange and a public commons already operate. Scarcity, who needs it? Managing it is is a bit like the iron lung industry, a second best that should be superseded by something better.