Peak paper

by John Quiggin on February 22, 2016

I’ve recently published a piece in Aeon, looking at the peak in global paper use, which occurred a couple of years ago, and arguing that this is an indication of a less resource-intensive future. Over the fold, a longer draft, with some links.

Since the dawn of history (literally, of written records), civilisation has depended critically on paper. As living standards have risen, so has the volume of paper produced, printed and read. The more knowledge we have and the wider its distribution, the more paper is needed.

At least, that was true until the end of the 20th century. With the rise of the Internet, the correlation between paper and information broke down. Increasingly, information is created and manipulated in electronic form, with paper serving mainly as an official record of the process.

In 2013, the world reached Peak Paper. World production and consumption of paper reached its maximum, flattened out, and is now falling. In fact, the peak in the traditional use of paper, for writing and printing, took place a few years earlier, but was offset for a while by continued growth in other uses, such as packaging and tissues.

China, by virtue of its size, rapid growth and middle-income status is the bellwether here; as China goes, so goes the world. Unsurprisingly in this light, China’s own peak year for paper use also occurred in 2013. Poorer countries, where universal literacy is only just arriving, are still increasing their use of paper, but even in these countries the peak is not far away.

The arrival of Peak Paper is of interest for a number of reasons.

  • First, it is, in large measure, the realisation of a prediction that was over-hyped in 20th century, and then derided in the early 2000s, namely, that of the Paperless Office.
  • Second, Peak Paper illustrates the meaninglessness of traditional concepts of economic growth in an information economy.
  • Third, the information economy that has produced Peak Paper implies a whole range, or mountainous terrain, of Peaks and Plateaus in natural resources of all kinds. Unlike the resource exhaustion scenario traditionally associated with the idea of Peak Oil, these peaks will be reached because improved living standards no longer require the ever-growing throughput of resources that characterized the 20th century industrial economy.

Let’s look first at the Paperless Office. The development of minicomputers and word processors in the 1970s led some farsighted thinkers to realise that computers would eventually have the same impact on office work, based on text, as they had already had on numerical tasks like payroll calculation. The phrase ‘the paperless office’ came to prominence in 1975, in a Businessweek article entitled The Office of the Future.

Initially, however, the rise of computers had the opposite effect. Computerisation generated vastly more information, which could be revised and reformatted in many different ways. But nearly everyone wanted to receive their information on paper, as what used to be called ‘hard copy’. The paperless Office of the Future appeared as a utopian vision, never to be realised.

The key point in the reaction was the publication of ‘The Myth of the Paperless Office’ by Abigail Sellen and Richard Harper, which crystallised the perception that the paperless office had failed. One wit suggested that ‘the paperless toilet will arrive before the paperless office.’

As it turned out, however, Sellen and Harper’s book appeared just as the paperless office was becoming a reality. As computer screens became more readable, and people learned to work with email and PDF documents, practices like printing out documents for offline reading declined. By now, we have reached the point where, far from being preferred, paper documents are subject to scanning and optical character recognition to get them into a digital form where they can be filed, searched and emailed.

The shift towards on-screen reading has affected other printed paper outlets, most notably newspapers and magazines. Surprisingly few newspapers have actually ceased publication, but nearly all have downsized, or even eliminated, their print versions.

Peak Paper is of interest to anyone concerned with the future of the world’s forests, but its significance goes well beyond that. Understanding Peak Paper tells us a great deal about the way the information economy of the 21st century differs from the 20th century industrial economy. Although the industrial economy is a thing of the past for most of us as far as work and daily life is concerned (In the entire United States, less than 2 million people are employed in large factories, and even adding China into the picture does not change things much), the conceptual categories of the 20th century still dominate our thinking.

Central to this thinking is the industrial model of economic growth, developed and formalised in the 20th century, and centred on the concept of Gross Domestic Product (GDP). The industrial model is one in which growth constitutes ‘more of everything’. More precisely, a growing stock of capital means that, using the same amount of labour, more and more resources can be processed into more and more final goods. This model leads naturally to the conclusion, central to the ‘Limits to Growth’ debate of the 1970s, that economic growth must eventually run up against constraints on the availability of natural resources, notably including trees to make paper.

A related, and critical, assumption, implicit in both the standard projections of ever-growing GDP? and in critiques like Limits to Growth, is that all sectors of the economy expand at a roughly equal rate. If this ‘fixed proportions’ assumption does not hold, the index number theory used to construct GDP numbers ceases to work, and the concept of a ‘rate of growth’ is no longer meaningful.

Peak Paper points up the meaningless of measures of economic growth in an information economy. Consider first the ‘fixed proportions’ assumption that resource inputs, economic outputs and the value of those outputs grow, broadly speaking in parallel. Until the end of the 20th century, these assumptions worked reasonably well for paper, books and newspapers, and the information they transmitted. The volume of information grew somewhat more rapidly than the economy as a whole, but not so rapidly as to undermine the notion of an aggregate rate of economic growth.

Throughout this period, the volume of printed books, newspapers and so on grew steadily, to around a million new books every year (Wikipedia). In total, Google estimates that 130 million different books have been published throughout history. The demand for paper for printing grew in line with that for books.

In the 21st century, these relationships have broken down. On the one hand, as we have already seen, the production of consumption and paper has slowed and declined. On the other hand, there has been an explosion in the production and distribution of information of all kinds. In 2010, Eric Schmidt of Google estimated that ‘Every 2 Days We Create As Much Information As We Did Up To 2003’. This claim has been the subject of some dispute, and is inevitably subject to definitional disputes. However, it is about the right order of magnitude if we compare the volume of digital information being created daily with the volume of information committed to paper throughout history

In any case, the estimate was out of date as soon it was made. The study on which Schmidt drew estimated an annual growth rate of 50 per cent in the volume of information being generated. Five years later, the volume is around seven times as large.

These estimates are consistent with personal experience. My first hard disk drive, 25 years ago, held 20 megabytes of data. My various storage systems now total about a terabyte, 50 000 times as much. That’s consistent with a growth rate of 50 per cent. Many readers will be able to confirm this for themselves by looking at their own data usage history and recalling that local storage has been replicated by ‘The Cloud” over the same period.

Finally, let’s consider the relationship between Peak Paper and the better known idea of Peak Oil. Information is now the primary engine of economic development and improved living standards, but we are still dealing with the legacy of the 19th and 20th centuries, when energy held centre stage. In particular, there is an urgent need to reduce, and ultimately eliminate, our use of carbon-based fuels in transport, industry and electricity generation.

There is a widespread belief that this goal can only be achieved with drastic reductions in living standards. Even in the absence of the imperative to decarbonise, advocates of the ‘Peak Oil’ hypothesis argue that an inevitable decline in the availability of oil will produce a sharp decline in living standards. This argument is another manifestation of the ‘fixed proportions’ assumption.

The analogy with Peak Paper shows why these beliefs are false. As with the historical relationship between paper and information, the demand for energy, and for fossil fuels to generate it, grew hand in hand with production of goods and services over most of the 19th and 20th centuries. And, as with paper, the industrial-era relationship between economic development and fossil fuels is no longer relevant.

The most notable example, all the more striking because it is central to so much misguided thinking, is that of oil. The world reached Peak Oil, in terms of consumption per person, in 1979. In the developed countries, the decline in oil consumption per person has outpaced population growth with the result that total consumption is declining. The average person in a developed country now uses less oil than their parents did 40 years ago.

This remarkable change hasn’t attracted much notice, for several reasons. First, much of the reduction in energy use has taken the virtually invisible form of improvements in energy efficiency. Both industrial processes and household appliances use far less energy than they did in the past. The only occasion on which this process has attracted any real attention has been with the ideological campaign by US Republicans to block the shift to more efficient lightbulbs, a policy that was legislated under the Bush Administration but implemented under Obama.

Second, until fairly recently, the main substitutes for oil have been other fossil fuels such as coal and gas. Oil-fired electricity generation was replaced by coal in the 1970s and 1980s, and then by gas. Oil for purposes such as home heating was replaced by a combination of gas and (fossil-fuel generated) electricity. It is only in the last ten years that renewable energy sources, most notably wind and solar photovoltaics, have begun to play a substantial role, but this trend is now well established.

Peak Coal has already arrived in the developed world. Coal consumption has fallen substantially in the United States and Europe, and is set to fall even further. Until recently, the decline in fossil fuel use in the developed world has been more than offset by rapid growth in developing countries, particularly in China and India.

But China has turned away from coal recently, largely because of the huge health costs associated with emissions of particulates and mercury. Beginning with Beijing, China has begun closing down all the coal-fired power stations located near major cities. Although construction of new coal-fired power stations has continued, it is no longer enough to offset the closure of older, dirtier plants. As a result of this trend, and a slowdown in the steel industry, China reached Peak Coal in 2013, at the same time as it reached Peak Paper.

India also is shifting the emphasis of its energy strategy to place more weight on renewables. The citizens of heavily polluted cities like Delhi are becoming increasingly unwilling to put up with the lethal effects of coal burning.

The case of Peak Steel is even more interesting. Steel lasts a long time can be recycled almost endlessly. On the other hand, but the need for steel demand is finite. In developed countries the stock of steel reached about eight tonnes per person decades ago, and has remained stable or declined slowly since then. With the stock of steel on a gently sloping plateau, the need for new steelmore can be met almost entirely by recycling scrap, in electric … rather than by burning coal to smelt iron ore in blast furnaces. So China has recapitulated the century-long development process of the West in the space of a couple of decades, rapidly becoming the world’s largest steel producer. The stock of steel has gone from near-zero to about 8 tonnes per person. While this stock will continue to grow for decades to come, the period of rapid growth is coming to an end. In terms of new production from blast furnaces, Peak Steel is imminent, too, if it has not already arrived.

All of these resources are, of course, used by people, and the number of people in the world is a crucial determinant. The rate of population growth has slowed substantially in the 21st century so far. Nevertheless, the central projections of the UN project continued growth for the rest of this century and beyond.

There are, however, good reasons for doubting these projections expecting an earlier stabilization of the world’s population. UN demographers are, for a variety of reasons, strongly attached to modelling approaches in which fertility rates for the world as a whole never fall below two surviving children per woman, the rate needed to replace a stable population over time. These approaches require projections in which developed countries, where fertility has been well below this replacement level for decades, experience an upsurge in fertility, and where the decline in fertility in the developing world is halted at or above replacement levels.

It seems more likely, on the whole, that the social trends that have reduced fertility in developed countries will be sustained in those countries and will extend more broadly throughout the world. If this happens, world population will stabilise around 2050, and decline slowly thereafter.
Peak Paper and the information revolution that produced it signal the end of the industrial economy that emerged in the 19th century and dominated the 20th. Peaks in resource use of all kinds can be expected to follow.

Contrary to what Peak Oil alarmists claim, these developments do not imply a reduction in living standards or an end to the process of economic development in countries that are currently poor. Rather, the information economy in which we are now living allows us to break the link between improving living standards and unsustainable growth in the extraction and consumption of material resources.

In the process, we have the chance to realise some of the most appealing aspects of the ‘degrowth’ idea proposed by the Club of Rome in the 1970s. The information economy allows a break withus to abandon the 20th century social model in which adults spend most of their days in an organized workplace. Instead, much of the value in the information economy is generated by informal interactions through various forms of social media. Combining this trend with steadily increasing productivity makes it possible to envisage a massive reduction in formal working hours, perhaps to the 15 hours a week envisaged by Keynes nearly a century ago.

On the other hand, the world of Peak Paper is not one the kibbutz archipelago of localism and self-sufficiency envisioned by the Club of Rome. Rather it is one in which people interact in many and complex ways, largely unconstrained by location. Its full implications will be hard to discern, however, until we break with the mental categories of the 20th century, and develop new ways of thinking about the information society.

{ 103 comments }

1

max 02.22.16 at 2:24 am

Contrary to what Peak Oil alarmists claim, these developments do not imply a reduction in living standards or an end to the process of economic development in countries that are currently poor.

Because they’re talking about a supply crunch and you’re talking about a fall in demand (demand crunch). A true supply crunch in oil would hurt very badly, but we’re very far away from that.

Its full implications will be hard to discern, however, until we break with the mental categories of the 20th century, and develop new ways of thinking about the information society.

I’m not sure the paperless office is entirely here yet (we’ll see), but I can definitely say we won’t know what the mental categories are that apply to the 21st century until after they occur.

max
[‘That is to say, we can’t dump the old model yet, for many reasons, such as the fact that humans are physical creatures. Upload everyone’s brains and that will change.’]

2

max 02.22.16 at 2:24 am

But a excellent post anyways!

max
[‘Sorry.’]

3

Rich Puchalsky 02.22.16 at 2:39 am

This piece is an example that illustrates how we’ve passed Peak Editing. I basically agree with it though.

4

nnyhav 02.22.16 at 2:40 am

Spoken like a true peakonomist

5

marku52 02.22.16 at 4:56 am

It seems to me that any work that really needs doing still requires energy. Is the information age going to take out my trash, or hoe my vegetables? Carry my produce over that hill?

So we are less wasteful in producing corporate memos. A fine thing but still “whoopy”. The real problem still remains. Your terabyte drive isn’t going to move any mass anywhere. Peak oil is still peak energy per capita.

6

Piers 02.22.16 at 5:00 am

“Since the dawn of history (literally, of written records), civilisation has depended critically on paper. As living standards have risen, so has the volume of paper produced, printed and read. The more knowledge we have and the wider its distribution, the more paper is needed.”

Pedantic book historians, like myself, might remind you that we have plenty of records from periods before paper displaced early media such as clay tablets, papyrus and parchment.

7

Matt 02.22.16 at 7:58 am

To make a tonne of pure silicon requires about 2.14 tonnes of silicon dioxide from sand, flint, chert, or quartzite and 0.86 tonnes of carbon from charcoal or coking coal. That is enough silicon to make 250 kilowatts-peak of PV modules of the more efficient crystalline silicon types. Based on a US average capacity factor of 25.9% for utility scale PV facilities and a 25 year lifetime*, that much PV capacity will generate:

365.25 * 24 * 250 * 0.259 = 567,599 kilowatt hours over its lifetime.

Based on an average 24 MJ/kg specific energy for coal and 1/3 thermal efficiency, it would take ((567599 * 3.6) / 24) / 0.33 / 1000 = 258 tonnes of coal burned as fuel to produce that much electricity in typical existing US coal plants.

Using carbon to make silicon PV devices is nearly 300 times more mass-efficient as a way to make electricity than directly burning the carbon as fuel. Of course that means redundancy for the vast majority of coal mining jobs. The job reduction from running PV plants vs. running coal plants is also severe; 10 people can monitor and maintain hundreds of megawatts of PV capacity.

The large increase in wind/solar “green jobs” is a dynamic effect borne of rapid growth for the sector in a developed-world electricity market that is flat or expanding very slowly. At equilibrium, the renewable energy sector will provide substantially lower employment per unit of energy output. If only the surplus workers’ mortgages and grocery bills would “dematerialize” too.

It’s a relief that reducing material intensity is both possible and a trend that is gaining force. That’s necessary, if not a guaranteed cure, to avert the worst predictions of resource shortage crises. But it’s also worrisome that the gains of dematerialization are not being distributed for the benefit of all. We’re not all going to make a living selling software and poetry to each other. Is it different this time? Or will my worries seem as quaint in a couple of decades as those who feared mass unemployment after farm mechanization?

8

Matt 02.22.16 at 8:08 am

Just after I posted I noticed an interesting parallel: my concern that green jobs are big right now only because they’re eating up a bigger, older fossil-energy world is similar to the observation that software is “eating the world”. Software developers are in high demand and paid well, but much of that strong demand is based on projects to reduce labor requirements in other sectors. In a better world it would mean at least the same comfort for everyone and fewer hours spent working. But so far we’re shrinking the work force instead of the 40 hour week.

9

Matt 02.22.16 at 8:17 am

Also I totally botched the math on my first post. 250 kilowatts of solar PV capacity will produce

25 * 365.25 * 24 * 250 * 0.259 = 14,189,962 kilowatt hours of electricity over its assumed 25 year lifetime*. Making silicon from carbon is thousands of times less mass-intensive than burning the carbon as fuel, not hundreds. This is offset somewhat by producing non-silicon components of PV facilities (steel for racking, copper wiring, concrete for anchoring), but not nearly enough for mining/manufacturing/plant employment at equilibrium to stay as high as it was when King Coal reigned over electricity.

*Median module output drops by about 0.5% per year, but field lifetimes for hardware are also usually higher than 25 years. I have neglected both of these small corrections.

10

Julian Bond 02.22.16 at 9:35 am

In the section near the end you talk about the UN population forecasts. And in particular “It seems more likely, on the whole, that the social trends that have reduced fertility in developed countries will be sustained in those countries and will extend more broadly throughout the world. If this happens, world population will stabilise around 2050, and decline slowly thereafter.”

For the last 4 decades total population has been growing linearly at 80m pa, 1b every 12-14 years. It’s not currently showing any signs of slowing so it seems likely that barring Black Swan events, it will continue like this for at least the next 15 years. Indeed, the UN revised their predictions last year for when we reach 10b, pulling it 6 years closer because their previous predictions of fertility rates falling weren’t happening as fast as expected.

2050 is now only 34 years away. So if world population is really going to peak by 2050, when will we start seeing a slow down in the linear growth?

11

John Kozak 02.22.16 at 11:04 am

The paperless toilet had been around for 20 years when “The Myth of the Paperless Office” was published…

12

Peter Erwin 02.22.16 at 12:47 pm

Peak Paper is of interest to anyone concerned with the future of the world’s forests …

Well, perhaps, though it’s a little ironic that you’re basing this on a report which is titled “World wood production up for fourth year”. So it’s not clear that “Peak Paper”, if we’re really there, is having much of an effect on the world’s forests.

Also, if you look at the most recent UN Food and Agriculture Organization report (as opposed to the previous year’s report you linked to), you’ll find the following:

Production of paper stagnated in Europe and declined in North America in 2014, but grew modestly in Africa, Latin America and Asia-Pacific.
After a slight decline in paper production in China in 2013, paper production and consumption resumed growth there last year, driving the overall growth trend in the Asia-Pacific.

So it may be a bit premature to declare Peak Paper in China.

13

magari 02.22.16 at 12:54 pm

John, is there a good or “standard” data set on world historical material throughput? Is The Weight of Nations considered reliable?

14

David 02.22.16 at 2:48 pm

The paperless office was a myth when I started working in offices in the 1970s and still largely is. In fact, the fall in demand for paper is probably much more indirectly linked to the IT revolution: things that used to be printed (like telephone books and airline timetables) have largely disappeared, consumption of newspapers and magazines has gone down and handbooks, guides, manuals etc. that used to be printed are now on CDROM, if they are physical at all. Ironically, the greatest saving is probably in IT itself: my first desktop computer also had a 20M hard disk (with MSDOS, WordPerfect and Lotus 1-2-3) but I had a shelf full of manuals to help me use them. The common point is that these innovations all have more advantages than drawbacks.
By contrast, the paperless office was always driven by accountants and IT people, rather than those who actually worked in offices. Electronic storage of documents, as well as reading them online, are subject to certain inherent limitations which can’t be overcome in the foreseeable future: thus the recent explosion of interest in analogue note-taking and old-fashioned technologies like the pen, which don’t need a power-supply and can’t easily be spied upon.
And what do you do in (to take a random example) a University Economics Department when a student is suspected of plagiarism and their thesis needs to be re-examined by a committee? Does everyone turn up with their own laptop with their own copy? “That bit’s not underlined in my version – oh yes, I see, it’s in light blue for some reason”. “No that whole sentence is underlined, Oh I see it’s because my copy of Word was set to German last week and I forgot to turn it back” “Sorry what mark did we just decide , I thought I’d saved the spreadsheet but I can’t find it….”
Technologies (including paper) will fall out of favour when they stop being useful, and not before.

15

ET 02.22.16 at 4:55 pm

@ David I suggest the University Economics Department committee use plagiarism checker software.

16

David 02.22.16 at 5:41 pm

@ET. They do, as far as I know. Certainly the (non-economic) university departments that I deal with do. Plagiarism software is an excellent example of the use of IT for a good and productive purpose, and, whilst it’s not always easy to use on screen, it makes it possible to do something that would be impossible any other way. But that’s not the point. If you want to get people round a table and crawl through through a text together, paper is your easiest and simplest option.

17

Matt 02.22.16 at 6:23 pm

I have worked for SF Bay Area tech companies for 10 years and the mostly-paperless office seems to be a widely accepted reality in this sector at least. Sketching notes and diagrams for yourself, sure, that still happens. It takes up less than 7 sheets a week in my case. There’s still a networked multifunction printer available but it’s used maybe once or twice a day across the whole office. Almost everything that would be printed 20 years ago is now viewed and annotated on a screen.

Everyone does bring their laptops to a meeting if we’re going over some shared document together. Technical manuals are all electronic. Even my employment agreement was delivered, signed, and returned as a scan without touching paper (I pasted an old scanned signature of mine onto it because I didn’t want to hook up a printer at home for the first time in months). If people want to read work material on the BART they do it with their smartphones. Even the old use case of “print it out to read when you don’t want to lug your laptop” is pretty much over.

By way of contrast, right before I started working in the SF tech scene I was employed at a national laboratory in the US. That job did use a lot of paper. I think I received an inch-thick stack of it on my first day. I probably printed 40 pages a day of charts and graphics just to show my research progress at each afternoon’s meeting. I don’t know if that organization has dematerialized in the years since or not. It might explain how different people have very different impressions of the paperless office’s progression.

18

Neville Morley 02.22.16 at 7:04 pm

Seconding Piers on the historical exactitude: paper in the strict sense – unless you’re using it in a very loose way to mean any sort of flat surface on which one writes – is a pretty late development. This does raise further questions about costs, accessibility and the like – papyrus and parchment were pretty expensive. But also note Roman use of wax tablet for short-term, disposal notes in easily reusable format – just have to melt wax to smooth it out and use for new memo.

19

Eric H 02.23.16 at 2:43 am

Excellent post! Too bad Julian Simon isn’t around; he could be publishing The Ultimate Resource III Kindle edition right about now.

In my previous job at a major DoD test range, we replaced film (I had my own personal Kodak rep for a while) with high speed digital imaging systems many years ago. As a result, I was convinced that a paperless office couldn’t be that far behind. The main factor keeping us in paper was the insistence by the gov’t that most forms still had to have wet ink signatures. We could not convince them at that time that a hashed signature was orders of magnitude more difficult to forge than a wet ink signature. In my current job, I rarely use a printer.

20

John Quiggin 02.23.16 at 4:26 am

Peter Erwin @12 You’re right that paper isn’t the only purpose for which trees are cut down. And as I noted in the OP, packaging and related uses are still rising. OTOH, the FAO report does support the central argument of the post

Production and consumption of graphic papers (newsprint and printing/writing) in the Asia-Pacific region has been decreasing since 2013, following the same trend towards digital media prevailing in America and Europe for more than a decade< ?blockquote>

21

Scott P. 02.23.16 at 4:46 am

“So we are less wasteful in producing corporate memos. A fine thing but still “whoopy”. The real problem still remains. Your terabyte drive isn’t going to move any mass anywhere. Peak oil is still peak energy per capita.”

Not as other sources come online. And by 2050 we’ll have fusion power, which will completely rewrite the script.

22

js. 02.23.16 at 4:48 am

what used to be called ‘hard copy’

Wait, isn’t it still called “hard copy”!?

(Sorry this isn’t very substantive, but it’s a little disturbing if “hard copy” has fallen out of use and I’ve just failed to notice.)

23

James Wimberley 02.23.16 at 7:55 am

Magari in #13: the latest paper by Thomas Wiedmann uses “material footprint” in tonnes and an I/O model including trade. Link in a post of mine at samefacts.com (I can’t get links to work from my smartphone, being ancient).

24

magari 02.23.16 at 10:51 am

Thanks James, I’ll look that up.

25

James Wimberley 02.23.16 at 11:09 am

JQ: “Peak Steel is imminent, too, if it has not already arrived.” It definitely has, if you can trust the statistics from the World Steel Association showing a peak in 2013. Chart and link to spreadsheet with the data in my post here. New iron from ore fell more (68 mT in 2 years, -5.3%) than steel. This is exactly what you would expect, as recycled steel from the rising volume of scrap metal uses less energy and is therefore cheaper. The substitution will continue.

26

Peter Erwin 02.23.16 at 12:49 pm

James Wimberley @25:
If I’m interpreting your figure correctly (unfortunately, I don’t have anything that can read .ods files), the disturbing thing is that that steel production has apparently become slightly less efficient in its use of recycled steel over time, because the proportion of steel produced from new iron has risen over time. The proportion of new iron in steel ranges from about 70-75% in the mid-1980s and 1990s t0 almost 80% in the 2010s. Maybe the last two years is a harbinger of what’s to come (we can hope), or maybe it’s a meaningless short-term fluctuation. (And I could point out that your figure shows “peak steel” happened in 1989, and then again around 1997, and again in 2007.)

27

Clay Shirky 02.23.16 at 1:09 pm

Contra @David #14 (and contra a lot of talk around that book when it came out), the relationship between information and paper was well and truly broken by the 1990s. The rise of two technologies — corporate databases, and office laser printers — collapsed paper’s older functions as a medium of composition, a platform for editing, a vehicle for transport, a surface for display, and a record for storage down to just one — display.

The percentage of written and visual information in the average office stored on paper went from nearly 100% in the early 1970s, to less than 1% by the late 1990s, which is to say that if anyone had printed every draft of every email, powerpoint, word processing doc, spreadsheet, and database, the volume of paper would be vastly larger than what was actually stored. The rise of the database (and it’s less well organized cousin, the hard drive) meant that electronic information was already, by the mid-1990s, growing at vastly much higher rates than any information stored physically.

This would have led quickly to the paperless office predicted by Business Week in 1975, but for the laser printer. The preference for changing fonts, printing out Powerpoint decks for meetings, and looking at designs on paper before adjusting them on the screen led to a massive increase of paper used per document. (Xerox once estimated that 5 pieces of paper out of 6 were only used on the day they were printed.)

Even as the percentage of information encoded on paper collapsed, the absolute volume (merely) doubled.

28

magari 02.23.16 at 1:28 pm

James, you seem to be omitting from your analysis the slowdown of the global economy, and in particular China’s economy, which surely must be behind the recent leveling off of material throughput in things like coal and steel. The big issue is whether we can decouple growth and throughput; we already know that economic stagnation reduces throughput.

29

Trader Joe 02.23.16 at 1:52 pm

An interesting post.

One thing I’ve wondered about is whether paper production is actually less energy intensive than perpetual maintenance of a comparable electronic document. By way of example, suppose suppose my business generates 1000 pages worth of information a day. If I were to print it, I could store it indefinitely for essentially no incremental usage of energy. Most of the energy usage would have been to proudce the paper in the first place.

By contrast, if I store it on my servers which need to be powered, and also cooled indefinitely in addition to requiring a perpetual steam of new servers on which to keep it (which usually involve an assortment of heavy and rare earth metals). Pound for pound over many years am I more green to print and retain (particularly using recycled paper) or more green to store digitally. Its also my observation that most people are quicker to actually discard and often recycle paper files whereas digital files are rarely purged and often maintained indefinitely.

I don’t know the answer to this, I suspect that paper usage isn’t as wasteful as it might at first seem but don’t really know enough about the process or energy consumption to know for sure.

30

Matt 02.23.16 at 6:32 pm

If I’m to believe this table from Low-Tech Magazine (primary sources are mostly paywalled), then new paper from timber has a midpoint embodied energy of 38 MJ/kg.

A 500 GB Samsung 840 EVO solid state drive — a typical choice for a nice laptop — masses 8.5 grams. Again according to Low-Tech Magazine (which is doing a bit of axe-grinding against electronics, but anyhow…) it can take 72 MJ to make 2 grams of microchips. Conservatively estimating the mass of the SSD as entirely microchips, that’s 306 MJ embodied energy, or as much as 8 kilograms of paper. 8 kilograms is about 1500 sheets of ordinary 8.5″ x 11″ laserjet paper. A graphics-heavy page of office printout might take up 4 megabytes of storage in digital form. Text alone like most email is 3 orders of magnitude smaller per page. There’s at least 125,000 page equivalents stored on the SSD for the same energy footprint as printing on paper once, or at least 83 times more information storage per unit of embodied energy.

You can count paper storage on-premises as “no incremental usage” only if you were going to occupy the same building footprint with or without the paper records, and you happen to fit the paper into space that would otherwise go bought but unused. Given the amount of floor space given up to filing cabinets in old pictures of offices, I’m not sure that is a safe assumption.

Likewise, storage of digital documents can be considered as having no incremental cost if you were going to buy the same storage capacity with or without document retention, and you’re just wasting less of it. That can happen too; even low end desktop computers come with hundreds of gigabytes of storage nowadays. That’s enough to hold a few lifetimes’ worth of email, memos, reports, presentations, spreadsheets, and other typical office fare.

Any well-run office is going to have offsite backup service for any digital documents that are still important. I don’t know if this was common for paper records back in the Paper Ages or if paper’s greater resistance to accidental deletion was considered adequate, and people didn’t worry about more severe accidents. Offsite backup protects against fire and water damage too, but I don’t know if those more extreme but infrequent risks were enough to spur offsite backups of paper records. In any case, based on the enormously greater storage density of digital media, I don’t think extra backups are enough to make digital storage more energy-intensive than paper.

31

James Wimberley 02.23.16 at 6:36 pm

Peter Erwin #26: there is only one maximum and it was in 2013. The others are just cyclical bumps. Tee local 2007 peak came just before the GFC, a classic shock. 2013-2014 were not years of global or Chinese recession. The recent decline in the proportion of recycled steel is surely just an artefact of recent high growth rates in output; the only way to get these is from new iron. It works the other way in a decline.
You can get LibreOffice for free, which I recoomend, and IIRC recent versions of Excel read .ods files.

Magari #28: what slowdown? When the Chinese say this, it means GDP growth has dropped just below 7%. It looks much, much worse from the stock market, because of the effects of high leverage on company profits. The world economy putters along at 3% or so growth. There is no recession.

32

Dan 02.23.16 at 6:54 pm

“One wit suggested that ‘the paperless toilet will arrive before the paperless office.’”

Not a wit exactly but Bob Metcalfe, one of the inventors of ethernet: “Wireless mobile computers will eventually be as common as today’s pipeless mobile bathrooms” Computerworld, Aug 16, 1993.

I thought it was pretty funny back then. Maybe not so much now.

33

David 02.23.16 at 7:35 pm

Clay@27. We are talking about different things. No one doubts the explosion of information since the 1970s and the fact that it”s almost entirely stored on electronic media. It could hardly be otherwise.
But this does’t mean that paper is going to vanish. There are, as I said, various technical and human reasons why this is so, and in some sectors (education and government for example) we have probably passed Peak Paperless, and the pendulum is swinging back the other way. Most people know that it’s easier and more effective to take notes in a notebook than on an iPad. I use technology for all sorts of things and always have, but I use paper where that makes more sense, which it often does.

34

Trader Joe 02.23.16 at 7:53 pm

@30 matt,

Neat analysis. Its a good point that to the extent data resides within chips they are also relatively costless to the extent that it doesn’t require incremental energy or cooling to maintain them.

I suppose part of the couterpoint however is the massive proliferation of electronic media which is maintained that might never exist but for the low cost to obtain and maintain – most everyone I know has gigabytes upon gigabytes of ‘cloud’ saved photos that are maintained in energy expensive server farms….they exist only because it “costs nothing” to take a picture on your phone and save it. Probably 99.99% of these photos wouldn’t exist if they had to be processed in the old way.

Your math proves pretty definitively that digital media is less energy intensive than paper for normal business usage – but I suppose some of that benefit is consumed by the sheer volumes of other digital media created and maintained.

Thanks for troubling to do my math for me.

35

Matt 02.23.16 at 8:53 pm

I suppose part of the couterpoint however is the massive proliferation of electronic media which is maintained that might never exist but for the low cost to obtain and maintain – most everyone I know has gigabytes upon gigabytes of ‘cloud’ saved photos that are maintained in energy expensive server farms….they exist only because it “costs nothing” to take a picture on your phone and save it. Probably 99.99% of these photos wouldn’t exist if they had to be processed in the old way.

Your math proves pretty definitively that digital media is less energy intensive than paper for normal business usage – but I suppose some of that benefit is consumed by the sheer volumes of other digital media created and maintained.

Yes, this is all true. It’s hard to get a handle on the real energy/environmental demands of cloud storage for at least 3 reasons:

1) Technology is still improving rapidly year-over-year per unit of storage. An analysis from 4 years ago is woefully out of date and biased toward over-reporting energy intensity per unit of storage.

2) “Hyperscale” data center operators — Google, Apple, Amazon, Microsoft, Facebook, and perhaps a few others — account for the lion’s of cloud storage and achieve power efficiencies significantly above ordinary multi-tenant data centers. But they also consider their energy-thrifty ways a competitive advantage and don’t publish detailed numbers. Using more accessible numbers from multi-tenant data centers also biases unit energy intensity upward.

3) There are vested interests at work trying to show that data centers need lots of electricity/fossils, and environmentally-minded folks are sometimes taken along for the ride, willing to believe the worst.

A particularly bad offender on point 3 is a coal-is-vital-and-sexy report called “The Cloud Begins With Coal,” a 2013 exercise in disinformation sponsored by the National Mining Association and the American Coalition for Clean Coal Electricity. It systematically overestimates energy demands of information technology and the necessity of coal for providing that energy. I can find this junk cited by news outlets as recently as July 2015.

Hyperscale data center operators have procured much more renewable energy for their operations since Greenpeace began pressuring them with its “How Clean is Your Cloud?” reports in 2012. In recent months I have seen multiple half-baked “debunkings” of this shift from right wing think tanks appearing on news sites.

36

magari 02.24.16 at 1:13 am

James, using China’s official GDP figure is naive. There’s a slough of stories with data suggesting big sectoral slumps. Considering China is the world’s largest consumer of these primary materials, if China slumps they demand less and ergo less material throughput. Copper is a good proxy for this one.

37

John Quiggin 02.24.16 at 4:20 am

@36 The problem, as pointed out in the OP, is that the converse is no longer true. A slump in China means less material throughput, but a decline in material throughput does not imply a slump. China, like most of the world, is a now a service economy.

http://blogs.wsj.com/economics/2015/10/23/china-pessimism-is-overblown-imf-says-citing-booming-services-sector/

38

Mark Bahner 02.24.16 at 4:42 am

If not “peak paper,” then “plateau paper”:

World cardboard and paper production

39

Mark Bahner 02.24.16 at 5:05 am

“A graphics-heavy page of office printout might take up 4 megabytes of storage in digital form.”

I deal with multi-hundred page documents with graphics every day that are 15-20 meg as Word or Acrobat files.

40

magari 02.24.16 at 11:12 am

@37 except that we’re seeing less materials consumed in the midst of a slump. We won’t know whether we’ve actually decoupled growth from material until economies start growing again.

41

PlutoniumKun 02.24.16 at 11:12 am

Nice to see an article that is a little bit optimistic. There is I think plenty of evidence that we are hitting ‘peak demand’ in a lot of products – cars (and miles driven per person), even house sizes. My suspicion is that a decline in use could be quite rapid – once investors realise a ‘peak’ has been reached there will be far less investment in supply, raising prices, so reinforcing the process. So much investment in base product production is done on an assumption of long term growth in demand – once that assumption is undermined, the investment climate changes entirely.

The Peak Oil issue is interesting and complex – I don’t think there is any doubt that we will see a major decrease soon in oil and gas production due to the massacre currently going on among producers. Common sense I think dictates oil consumption will reduce per person – I can remember as a child the huge tank of oil needed to keep our house warm for the winter (greatly reduced when we insulated and double glazed the house), and how inefficient my fathers old Chrysler was. Growth in wealth around the world disguised this process.

As for Peak Paper – in the late 90’s I worked for a major engineering company which aspired to go paperless. This was a source of some hilarity as the new photocopier machines and printers were used to churn out vast amounts of worthless copies of every memo, minute and plan produced. Fat fingers errors used up mountains of paper as someone would accidentally print a 500 page contract document instead of the single page they wanted. But slowly but surely I have seen how the size of files is reducing as more and more stuff goes online. In my job, what was once a paper process (writing out to a local authority or company for copies of legal documents) is now bypassed because I can just click onto their website and look at whats there.

42

TM 02.24.16 at 12:27 pm

I don’t know about paper in particular but generally speaking, decoupling of economic growth from material consumption isn’t happening yet. The resource intensity of economic activity is by most measures in decline but nowhere near fast enough to compensate for doubling of the economy every 20 or so years. Some time ago I crunched the numbers on energy consumption. Intensity declines by at best 1% per year (can’t look up the details now).

As to the wider question of whether improving living standards are consistent with declining resource use: The answer is yes, up to a point. But first we need to agree that GDP is not just a lousy but a totally irrelevant measure of living standards. GDP doesn’t measure most factors that we associate with standard of lving, and those that it correlates with can be better measured directly. Economists have theoretically known this forever yet they simply can’t get the obsession with economic growth out of their heads. So assuming that we have a better metric of living standard or quality of living, and supposing that everybody in the world has essentially reached West European standards on that metric, is it really plausible that that metric will just go on increasing?

I think it’s not. The ultimate limit is time.

We can only do so many things in a day. For example, we could listen to millions of music pieces and watch millions of movies, at very little economic and resource cost, but we don’t have the time. Technical progress can reduce the necessary work time and thus increase the amount of leisure time we enjoy, but only within the limits of the 24 hour day. And it is arguable whether the internet and smartphones have made our use of time more or less efficient – I suspect it’s the latter.

43

JW Mason 02.24.16 at 1:04 pm

Central to this thinking is the industrial model of economic growth, developed and formalised in the 20th century, and centred on the concept of Gross Domestic Product (GDP).

I’m confused by this sentence and the following few paragraphs. Are you saying you think the distribution of activity across industries or sectors within a given country will change much more between, say, 2000 and 2050 than it did between 1950 and 2000, or 1900 and 1950, or 1850 and 1900, or 1800 and 1850? And, are you saying that GDP comparisons between, say, 1950 and 2000 implicitly assume there was no change in the mix of output during this period and are meaningless insofar as there was? The first statement seems very strong — even if “information” is big, is it really bigger than industrial revolution? — and the second simply wrong. But I suspect I’m misunderstanding you.

44

JW Mason 02.24.16 at 1:08 pm

Also, let’s not confuse the national accounting conventions behind GDP with Y = f(K,L). These two conceptions of growth get kludged together in various ways but they really inhabit separate conceptual universes. GDP as such doesn’t depend in any way on the idea of a growing stock of capital.

45

Daniel 02.24.16 at 2:40 pm

Yes I’m with Josh on this one – GDP as a concept is a sum of final transactions; it’s not an index number. Do you mean the deflator (and thence the concept of real GDP I suppose)?

46

James Wimberley 02.24.16 at 5:48 pm

TM 342: “Some time ago I crunched the numbers on energy consumption. Intensity declines by at best 1% per year ” Wiedmann calculates that material footprint in tonnes has been growing at 60% of the rate of GDP. If the latter is 4%, the decline is 1.6%, which is not too far of your energy indicator. (BTW, why worry about massless energy?) Your problem is the spate on recent indicators suggesting an inflection. I counted four without trying very hard: paper, iron and steel, coal, and carbon emissions.

Magari #40: you will have to do better to discredit China’s GDP numbers, endorsed by the IMF, and convince us of your imagined “slump”.

When national statistics agencies were set up, advanced economies were dominated by manufacturing,mining, and agriculture – the material sectors -, so that’s what they concentrated on. The bias is maintained by the ease of counting: bushels of wheat, barrels of oil, ounces of gold, tonnes of steel, numbers of cars and washing-machines. The exercise can be extended to orderly, centralised services like insurance. But you try extracting accurate and timely reports from plumbers, car repair shops, bars and dentists: they don’t cooperate and if forced, tell lies.

47

Mark Bahner 02.24.16 at 6:09 pm

“Some time ago I crunched the numbers on energy consumption. Intensity declines by at best 1% per year (can’t look up the details now).”

As far as I know, in the U.S., primary energy consumption is still a bit below the 2007 peak:

U.S. primary energy consumption

48

John Quiggin 02.25.16 at 1:32 am

The crucial concept in national accounts is value-added. The key idea is to follow a physical good through all the primary, secondary and tertiary stages: the primary production of the raw materials, the manufacture of the good in the secondary sector, and its transport, distribution and sale to final consumers, without double counting.

This way of looking at the economy makes sense specifically for a 20th century industrial economy. It is useless in an economy based on subsistence agriculture and unhelpful in an economy based on information or pure (not associated with delivery of a good) services.

49

magari 02.25.16 at 1:53 am

James, are you joking? There’s been an avalanche of such stories in the past couple of years. Here’s one to whet your appetite: http://www.wsj.com/articles/chinas-true-growth-is-a-mysteryeconomists-weigh-the-clues-1430071125

50

Peter Erwin 02.25.16 at 10:00 am

John, I’m not objecting to the idea of Peak Paper, or even that we’ve already reached it. For example, the stagnation in paper production and use in advanced economies is actually very suggestive, and I think you would have done better to focus on that rather than promoting a single year’s decline as somehow definitive evidence that, yup, we already reached and passed Peak Paper. (It’s not as silly as the past tendency of global-warming denialists to look at a single-year’s decline in global temperatures — e.g., 1999 or 2000 — and declare, “Look — the warming is over!”, but it’s far enough away, either.)

In fact, by doing so you missed the opportunity to suggest that Peak Paper probably happened in 2007. If you look at the UN Food & Agriculture forestry data (which extends from 1961 through 2014), it shows that world production of “Printing + Writing Paper” actually peaked in 2007 and has declined since then. (There was a temporary recovery in 2010 — though never above the level of 2008 — but then the decline continues.) The fact that this pattern is not replicated for other wood products suggests that this not just the lingering effects of the Great Recession, and may indeed be related to the (belated) arrival of the paperless office.

Note, however, that this doesn’t apply to all paper products. “Household+Sanitary Paper”, for example, shows uninterrupted increases for every year since 1993, even during the Great Recession. (So, yeah, the paperless toilet hasn’t arrived yet.)

51

magari 02.25.16 at 10:59 am

Peter, indeed, the fact that the world has yet to adopt the Toto Washlet tells us that any hope we’ve decoupled material throughput from growth is wishful thinking! I jest. But if we think of the sheer magnitude of consumer products as well as transformations of the built environment that will have to take place for the developing world to match the affluence of the developed, then we get a sense of the enormous amount of the Earth’s resources that remains to be utilized.

52

TM 02.25.16 at 1:28 pm

47: I was referring to the world wide numbers. Energy use in developed countries has indeed declined in absolute terms, which is good news, but let’s not forget that some of this decline is not due to increased efficiency but to deindustrialization and outsourcing. If embodied energy (or carbon emissions) of imports are properly accounted for, the statistics look less pretty.

In any case, globally, energy use and cabon emissions are still growing, and they need to go down sharply. We are nowhere close to sustainability, yet.

53

TM 02.25.16 at 1:30 pm

50: Would be curious how recycled paper is accounted for in those stats.

54

Peter Erwin 02.25.16 at 2:13 pm

TM @53:
There’s a separate entry for “Recovered Paper”, which shows a pretty steady increase over time (only a tiny downward blip in 2012, and no decrease in 2008 or 2009).

Of course, whether the “Printing+Writing Paper” is coming primarily from wood pulp or from recycling is orthogonal to the question of whether office paper use is declining.

(Newsprint paper production, for what it’s worth, peaked in 2004.)

55

Mark Bahner 02.25.16 at 5:44 pm

52: “In any case, globally, energy use and carbon emissions are still growing,…”

That’s an assertion worthy of debate. Even if global energy use and carbon emissions are still “growing”, they’re not growing very fast:

BP on global energy use, 1989 to 2014

Enerdata on global energy use, 2000 to 2014

Similarly, by some accounts global CO2 emissions were flat or declined in 2014, and declined in 2015.

56

Julian Bond 02.25.16 at 6:37 pm

Hooray! Global CO2 emissions per year are only as high as they’ve ever been and not higher!

Meanwhile the CO2 in the atmosphere keeps growing.

57

TM 02.25.16 at 6:48 pm

55: I haven’t looked up the latest numbers. There was a decline in global emissions due to the 2008 recession but emissions were growing again the very next year, as I recall. According to your second link, China’s energy use has for the first time not grown. That would be an important milestone if correct.

58

James Wimberley 02.26.16 at 1:11 am

Julian Bond #56: As Will Rogers said, the first thing to do at the bottom of a hole is to stop digging.

59

Peter T 02.26.16 at 3:37 am

For those interested, I just came across Geoffrey Parker’s Global Crisis, on the effects of the couple of centuries of climate downturn around the C15/16 (some estimates are that human populations declined by one-third globally). As usual, the impacts are not straightforward, direct or uniform, but Japan seems to have come through better than most (which also seems to be the case in our time).

60

Garrett Wollman 02.26.16 at 4:34 am

Forgive me for joining the conversation very late indeed, but I wanted to correct a few misconceptions in the comments above regarding solid-state storage:

(1) magnetic disk (rotating) storage is still vastly cheaper than solid-state storage in cost per unit data, although it is more matter-intensive (the control electronics for a rotating drive account for about as much mass as the entire package of a solid-state drive), and there is no evidence that this is likely to change any time soon. Magnetic tape is cheaper still. This is in large part because the manufacturing facilities for magnetic media are vastly less capital-intensive than semiconductor fab lines (by orders of magnitude). Brent Welch at Google noted in a recent presentation that “World-wide manufacturing capacity of semi-conductor bits is perhaps 1% the capacity of making magnetic bits.”

(2) the time it takes to transfer the entire contents of a drive has been steadily lengthening — the drive-computer interfaces have not been getting faster at the same rate as drives have been getting more dense — which means that data must be spread out over more disks to allow for both performance and fault-tolerance. This means that 1 TB of “user data” stored in a cloud provider may actually be spread across a dozen different drives and consume substantially more storage after erasure coding or other forms of redundancy are taken into account.

(3) solid-state storage degrades while powered off (this is physical limitation, inherent in the technology — flash memory — which is currently used for such devices). These devices also have a much more limited lifetime, in terms of update cycles, than magnetic media — again a consequence of semiconductor physics.

(4) Those “exascale” data centers contain hundreds of thousands of rotating disks. These devices wear out over time — in fact, many fail every single day in a data center of this scale — but so far operators have been able to replace failed disk drives with newer models that have a larger capacity for about the same cost in the same physical footprint. It isn’t clear whether or how long this trend can continue, as it depends on ever-increasing areal density of magnetic media, and that is subject to some physical limits as well. Drive manufacturers have a “roadmap” that suggests 60 TB will be feasible in 10 years (that would be a sixfold increase over the current maximum capacity of individual spinning disks).

(5) The typical operating power of a standard drive is between 5 and 10 W, which makes drives a substantial fraction of the 50+ MW consumed by a large data center, with a concomitant requirement for cooling. This power requirement is determined mainly by the physics of the moving parts. Solid-state drives require less power, but the increased capital cost and limited durability limit their utilization to applications that do not require high density.

61

Mark Bahner 02.26.16 at 4:38 am

“Meanwhile the CO2 in the atmosphere keeps growing.”

Yes, over the 5 years ending in 2015, the CO2 concentration increased by an average of 2.3 ppm per year.

atmospheric CO2 concentration

And in 2015, the atmospheric concentration of CO2 reached 400 ppm. If the atmospheric concentration continues to increase by 2.3 ppm per year, it will cross 560 ppm (double the pre-industrial concentration of 280 ppm) in 2085, and reach approximately 600 ppm by 2100. Of course, it’s more likely that emissions will peak and go down, rather than continuing at the present rate.

That’s why, not coincidentally, a decade ago I predicted that the most likely CO2 concentration in 2100 was approximately 560 ppm, with a 90 percent chance that the concentration would be between 440 ppm and 635 ppm.

In contrast, a Wigley and Raper paper published in Science after the IPCC TAR was published predicted that the most likely CO2 concentration in 2100 was 720 ppm, with a 90 percent chance that the concentration would be between 570 ppm and 935 ppm. In other words, the paper published in Science predicted that there was a 95% chance that the CO2 concentration in 2100 would be more than 570 ppm.

It’s anyone’s guess whether Science will ever publish a more…scientific paper on the issue. (To do so, Science would have to actually be interested in science.)

2006 predictions

62

TM 02.26.16 at 8:59 am

I hoped for some more discussion on the question of decoupling. My view is that sustainable economic growth through decoupling (or qualitative growth) is not plausible long term. Paper books is a good case study.

A few hundred years ago, most people couldn’t afford to buy more than a few books a year if at all. So the number of books printed every year was small. Over time, economic growth made people richer so they could afford more books. Thanks to technical innovation, more books could be printed cheaper. Literacy vastly expanded and also the circle of people with the education and economic means to write books expanded. So more books were printed every year and also more paper was needed and more trees were cut. People began to worry that if the number of books printed every year continued to grow exponentially, we would soon run out of trees. More efficient production processes and recycling reduced the pressure on forests but that could only delay the point at which further growth would run into fundamental ecological constraints.

Technology has changed the equation. Printed paper is just a medium for exchanging information. We now have means to exchange information electronically without paper. In principle, we can now create a virtually unlimited number of copies of books with little (though not zero – energy and physical infrastructure are still needed) resource footprint. This, JQ and many others argue, means that “sustainable” economic growth can continue forever. The number of electronic books in circulation can grow almost without limit. If the book example can serve as a model for the economy in total, then by extension economic growth is not limited by resource constraints.

The obvious problem is that the number of books a person can enjoy per unit of time is limited. Even if gazillions of books are electronically “printed”, humanity won’t have the time to read and enjoy them. It would be absurd to count these virtual books as “economic growth” if there is no actual demand for them. The same logic applies to every industry that can plausibly be construed as an example of decoupled, qualitative growth: music download, videos streaming, telecommunications in general, but also wellness, massages, fitness, education, cultural events, and so on. The argument of decoupling is essentially that once our material needs are fulfilled, we can spend more money on “quality time”, living the good life, and this doesn’t necessarily involve a high resource use. I’m not dismissing the concept of qualitative growth per se, in fact I agree that it would be desirable to be able to engage more in these types of activities (although again it must be understood that they do involve a non-zero resource footprint). But all of these activities are limited by time and therefore growth in these areas will never plausibly make up for the limits of material, quantitative growth.

63

James Wimberley 02.26.16 at 12:43 pm

TM #62: So what? The thing we have to worry about is material consumption, and ultimately ecosystem footprint. If these are sustainably flat, it doesn’t matter much how we describe the expansion in immaterial information or energy and the immaterial component in services (like the skill in cooking applied to the material ingredients). Intuitively,it seems plausible that the price system invented to handle material scarcity will gradually break down and the socialist (public provision, much of it unpriced) and communist (unpriced sharing) sectors will expand. The blogosphere is good example of the communist sector. It is expanding, in page numbers and I suppose page views, it provides a valuable service, but it falls outside the GDP metric. JQ points out that the information society makes this metric less and less relevant. Interesting times.

Politicians like Cameron wittering on about the priority of growth should be asked insistently: growth in what? Per Piketty, when GDP growth is highly unequal, and the material wants of billionaires are sated, a growing part goes into capital accumulation. I don’t understand why he does not see a doomsday machine here, such as Keynes’ asphyxiation of the rentier by a falling rate of profit.

64

TM 02.26.16 at 4:10 pm

So what? Well most economists (including JQ if I’m not totally mistaken) and politicians are deeply committed to “the priority of growth”. The (im)possibility of sustainable growth seems a big deal to them.

65

Matt 02.26.16 at 7:31 pm

What economists and politicians might imagine about growth now, growth forever is not that interesting. It’s obviously incorrect in the long run but it doesn’t really need to be counter-attacked as an idea because consumption per capita will hit limits of either supply (old growth timber, wild cod) or demand (electronic books, salt, vodka…) regardless of belief.

More interesting to me are the actual material developments of the later 21st century: population, material resources and consumption per capita, overtaxed natural sinks for human waste products (most prominently CO2 but also many others), the neck-and-neck race between natural resource depletion/damage and improved processes of production. Can the people of Bangladesh reach the same high level of human development as the people of Costa Rica? Can people currently living in high-HDI nations maintain those indicators while decarbonizing and enduring the consequences of past failures to decarbonize? What are the most humane and just ways to sunset systems of production and consumption that are environmentally egregious (e.g. how to help ex- coal industry workers as coal declines)?

66

magari 02.27.16 at 2:50 am

Matt, yep, those are the interesting questions alright. I’m surprised at how little research is being done into them.

67

Metatone 02.27.16 at 8:53 am

@TM

I think you are connecting two separate issues.
Can we decouple? & What are the consequences of decoupling?
I’d go with the idea that we can decouple, but that the consequences for our current economic system are profound. There’s an inevitable limit to demand, complicated (as we see now) by the fact that there can be great needs amongst a population who don’t have money, so it doesn’t translate into demand.

@David – I hear what you are saying. I think there’s a misunderstanding about Sellen & Harper. Their key research wasn’t about global paper usage, it was about progress to a truly paperless work environment. Their key point was that technology did not as yet provide the “affordances” of paper. As such, computers did not enable all the thought processes that paper did. Having used the latest iPad and pencil a bit, I think we’re on the verge of changing that, but it’s important to recognise that we’re not there yet.

It’s true that in the meantime we’ve made good progress in general and at the same time invented a bunch of industries (largely centred on coding) where paper is used much less. However, there are still a bunch of electronic thinking aids yet to be perfected. (One might also consider the studies that show how the lack of particular affordances leads to particular design models (often fragile ones) being used in the coding industries…)

68

Peter T 02.27.16 at 10:26 am

If we end up with most people using paper notepads for shopping lists, stray thoughts and doodles, but all commercial transactions electronic, then paper consumption will be miniscule.

My guess would be that a rapid transition to public and unpriced (communistic) provision will be profoundly destabisiling, most of all to established elites. The trend of the last several decades has been to frantically shift previously public or unpriced activities into the priced sector to compensate as the “natural” size of the latter has shrunk.

69

TM 02.27.16 at 10:23 pm

65 “What economists and politicians might imagine about growth now, growth forever is not that interesting.”

Well it may not be “interesting” but it profoundly influences public policy. I would also argue that the near universal fairy tale belief in growth forever severely limits our collective imagination and therefore our future options.

70

pnee 02.27.16 at 10:48 pm

@62. Humanity doesn’t read all the same books or listen to all the same songs.

As the material cost of “manufacturing” a book or record drops to zero, more of the price of media can go to authors, who in turn would need fewer readers/listeners to have a work be a success financially. At the consumer end, the potential increase in choice could be considered an improvement in standard of living.

One can already see this in author published e-books; but an even better example may be “Peak TV” the idea that they are more producers of quality television shows than ever before, even though the individual shows have small audiences compared to the typical audience of the “Golden Age” of Television in the 1970s. This may be an artifact of a transitional period where broadcasters and streaming services are both striving to become (or remain) viable producers of new content. But I think that technology and a proliferation of production skills and creative talent has made TV production of larger numbers of programs for small, well-targeted audiences the “new normal”.

Obviously there’s still a limit. We can’t all be authors of e-books read by two or three of closest friends and no one else. But taking an expensive physical medium out of the equation takes some of the pressure off. Larger numbers of aritsits with more modest success is now viable, if no books had to be printed, no LPs pressed, and no film prints sent to theaters or TV stations.

71

John Quiggin 02.27.16 at 11:21 pm

@TM

I don’t see how you can read a commitment to “the priority of growth” into a post about “the meaninglessness of traditional concepts of economic growth”

72

TM 02.27.16 at 11:35 pm

Fair enough. I thought, but maybe misremembered, that you had expressed faith in continued economic growth in some earlier writings. What is your view then on my argument at 62?

73

Brett Dunbar 02.28.16 at 9:39 am

Some of the carbon excess will have to be dealt with by increasing natural carbon absorption. The only land ecosystem that acts as an ongoing carbon sink is peat bogs. Carbon is absorbed and laid down in layers of peat and they will absorb carbon continually for thousands of years. Britain and Ireland can do a bit by restoring wetland, ending peat cutting removing trees, blocking drainage ditches &c. Forests on the other hand are more a carbon store than a sink they absorb at a fairly high rate until mature and then the releases due to decay and fire matches the absorbtion rate. Marine ecosystems that produce carbonate rocks also permanently absorb carbon.

74

Tallmarc 02.28.16 at 2:11 pm

I found this conversation very interesting and just the tiniest bit hopeful. So much so that I printed the whole thing to read later.

Slightly more seriously, the possibility of decoupling of energy/resource/mass use from “economic growth”, however measured, is new to me. Equating economic growth with standard of living has been accepted dogma in my limited world (academic biology) for some time. Since so many people seek to achieve US/UK standards, and since so many lack basic human needs for food/water/health, the hope raised by this decoupling trend is too little, too late for humanity.

75

Brett Dunbar 02.28.16 at 6:15 pm

The energy intensity per unit or production has been falling continuously for a very long time. However total production has generally risen faster than the intensity has fallen. Still total energy consumption per capita in the UK and USA peaked in the 1970s (UK 1973, USA 1977) and has been more or less flat since then.

76

John Quiggin 02.28.16 at 11:36 pm

TM @62 To spell out your argument as I understand it, we can reach satiation in anything provided by information technology, but that doesn’t help us in relation to those things that require material throughput, which can’t be increased sustainably

First, the satiation argument also applies to lots of items of material consumption, and in fact satiation has been reached, or is close, for quite a few of them among the well-off in developed countries. I mention steel in the post – rich countries already have all the steel they will ever need: they just need to replace the small fraction of scrap that can’t be recycled. Food, clean water, car travel are more examples.

More generally, as I’ve argued in previous Aeon articles, I regard the living standards now obtained by the well-off in developed countries as being close enough to satiation that we ought to be focusing on increasing leisure rather than increasing consumption.

I’ve also argued that we can sustainably raise the living standards of people in poor countries to those enjoyed in the developed world now. Obviously that requires some big reductions in the materials throughput per unit of final services but that’s feasible.

Just summing up, I’m for satiation at a high level, not endless growth.

77

Plume 02.29.16 at 12:38 am

JQ,

The obvious trap of raising the bottom up, if we don’t radically lower the top down, is this:

The World Wildlife Fund estimates that by 2030, we’ll need two entire earths to meet demands for natural resources. And if the world lived like middle class Americans, we’d need four. There is no way out of this disaster if we stay with the current economic system. And that’s even if a miracle were to happen wherein the capitalist class would suddenly decide to work together for our benefit, for the planet’s benefit, for the benefit of the poor and the working poor, etc. etc.

Believing they will is the real “utopian” dream. Aint never gonna happen. But even if they suddenly all got together and acted as if they were in a Star Trek episode, it still wouldn’t be enough.

78

Mark Bahner 02.29.16 at 2:35 am

“The World Wildlife Fund estimates that by 2030, we’ll need two entire earths to meet demands for natural resources.”

Estimating the “footprint” of humanity isn’t science. It has no predictive power and isn’t falsifiable. We have only one earth.

79

Mark Bahner 02.29.16 at 2:54 am

“Can the people of Bangladesh reach the same high level of human development as the people of Costa Rica?”

I predict the people of Bangladesh in the year 2100 will have a higher per-capita GDP (expressed in year 2016 dollars) than the people of the U.S. today.

80

Peter T 02.29.16 at 3:06 am

Well, it’s falsifiable in the same way the statement “if you drink a quart of cyanide, you will die” is. You can disprove it by drinking a quart of cyanide and living. Would you recommend the experiment?

81

Plume 02.29.16 at 3:25 am

Mark,

“We have only one earth.”

Yes, and obviously that’s the entire point. We are destroying it, depleting its natural resources, polluting its atmosphere, land and water, and soon enough it won’t be able to sustain human life. Or much of any other kind, either.

In the last 40 years alone, we’ve literally lost half our wildlife and 90% of our fish stocks. That’s not a prediction. That’s actually happened. So these environmentalists, geologists, zoologists and overall scienzy guys can pull together data like that . . . . merge it with data on human consumption patterns, population growth and so on, and come up with some pretty good predictions. Pretty good in the “if you keep on doing X, then Y will occur” way. And the history of that kind of prediction over the last few decades has actually been quite “conservative.” Climate change is worse than was predicted even a few years ago. Pollution is worse. Species extinction is worse.

But if you want to claim we can’t possibly predict these things, so we should just fly blindly over the ecological cliff, I truly hope you’re in no position to make major decisions on these matters.

82

TM 02.29.16 at 9:54 am

76:

“I regard the living standards now obtained by the well-off in developed countries as being close enough to satiation that we ought to be focusing on increasing leisure rather than increasing consumption.

I’ve also argued that we can sustainably raise the living standards of people in poor countries to those enjoyed in the developed world now. Obviously that requires some big reductions in the materials throughput per unit of final services but that’s feasible.

Just summing up, I’m for satiation at a high level, not endless growth.

I fully agree. Thanks for clarifying.

83

TM 02.29.16 at 10:09 am

[Darn blockquote]

Mark 79: Banglasdesh GDP per capita has roughly tripled over the past 40 years. You are predicting a 70 fold increase over the next 84 years. I’m glad you are limiting yourself to strictly scientific predictions ;-)

84

Mark Bahner 02.29.16 at 5:20 pm

“Well, it’s falsifiable in the same way the statement ‘if you drink a quart of cyanide, you will die’ is. “

Writing that if things go as expected, the world will be using the equivalent of two earths in 2030 is not a falsifiable statement. Notice that they don’t predict any falsifiable consequence of the alleged two earths being used up. They don’t predict that the world life expectancy at birth will decline, for instance. Or that oil will be $200 a barrel, or copper $5 a pound. Or that the world economy will shrink or stop growing. Or that the cost of food will increase by a factor of two or ten. Or that X percent of species on the IUCN Red List will become extinct.

Writing that the global footprint will be two earths is essentially the same as wearing a posterboard that says, “Repent your sins!”

85

Mark Bahner 02.29.16 at 5:32 pm

“In the last 40 years alone, we’ve literally lost half our wildlife and 90% of our fish stocks. That’s not a prediction. That’s actually happened. So these environmentalists, geologists, zoologists and overall scienzy guys…”

“Scienzy guys.” Right. Sort of like “truthy guys.”

Here’s what Stuart Pimm at Duke says about their “science” in a National Geographic review of the “50% of all wildlife lost” claim:

Mix apples, oranges, grapes and peanuts…and divide by four

“But Pimm is skeptical of the approach WWF used to calculate the species loss. ‘I’m not a fan of this planetary index because it mixes a lot of different numbers together in an essentially arbitrary way-therefore, it’s hard to know what exactly is meant by a 50 percent loss of vertebrates over the last 40 years.'”

“For instance, you can’t put British songbirds in the same category as West African lions—’it’s an apples and oranges and pears and grapes and cookies index that lumps a whole bunch of things together in a way that requires a lot of effort to dissect all the different pieces,’ he said.”

86

Plume 02.29.16 at 5:36 pm

Mark @84,

This is, to be generous, a very poorly worded paraphrase of what the WWF is saying:

“Writing that if things go as expected, the world will be using the equivalent of two earths in 2030 is not a falsifiable statement. “

The World Wildlife Fund says we will need two entire earths to sustain our current lifestyle — four earths, if the entire globe lives like a middle class American on average. Obviously, this isn’t possible. Which is the entire point. They’re not saying that we will be using two entire earths — obviously. They’re saying that we would need two entire earths just to meet the demands of our current lifestyles . . .

http://wwf.panda.org/

http://www.worldwildlife.org/

87

Plume 02.29.16 at 5:40 pm

Mark @85,

Oh, come on. I chose the words “scienzy guys,” as a riff off of the movie, The Martian. The WWF didn’t.

You are really desperate to dismiss what the majority of scientists say is happening. Going after my wording is the last refuge of climate denialists, and it’s quite similar to gun fetishists who pounce on the use of the term, “assault weapons,” as if the language alone ends all debate in their favor.

88

Mark Bahner 02.29.16 at 5:41 pm

“Mark 79: Banglasdesh GDP per capita has roughly tripled over the past 40 years. You are predicting a 70 fold increase over the next 84 years. I’m glad you are limiting yourself to strictly scientific predictions ;-)”

A few questions:

1) Why do you think that looking at Bangladesh’s GDP over the past 40 years allows you to assess the science (or lack thereof) behind predictions for the next 85 years?

2) What factors do you think will affect Bangladesh’s GDP growth over the next 85 years?

3) Do you think this paper is not scientific?

Economic growth given machine intelligence

89

TM 02.29.16 at 9:37 pm

I’ll attempt a brief answer.
1) There are no properly speaking scientific “predictions” of nondeterministic processes. Physicists can predict the movement of planets and they can predict the climatic consequences of CO2 accumulation in the atmosphere but there is no science that can predict social processes. There are just more or less informed guesses. Looking at past trends is one useful source of such information. It is of course not permissible to simply extrapolate from past trends. But if you wish to predict a huge change in such a trend, surely the burden is on you to provide reasons why such a change should occur.

2) Many factors, and many of them are unpredictable. Among the more predictable factors are some that are very likely to have adverse effects. Climate Change is a factor which under most plausible scenarios is certain to have very negative, if not catastrophic, effects especially on Bangladesh (even more frequent flooding, water scarcity, etc.) I’m not aware of any plausible reason to believe that Bangladesh’s economy will perform much better in the next decades than it has in past decades, and there are strong reasons for assuming the opposite.

3) I briefly perused the paper. It derives certain results from a certain mathematical model under certain assumptions. I’ll assume (not having the time to check) that the treatment of the model is mathematically correct. What the paper doesn’t offer is any evidence at all for believing that the model correctly predicts real world economic processes.

90

Mark Bahner 03.01.16 at 4:16 am

“1) There are no properly speaking scientific “predictions” of nondeterministic processes. Physicists can predict the movement of planets and they can predict the climatic consequences of CO2 accumulation in the atmosphere but there is no science that can predict social processes.”

So no one can ever make a scientific prediction of future economic growth? If so, why did you imply my prediction of economic growth in Bangladesh was somehow unusual because it was unscientific?

“2) I’m not aware of any plausible reason to believe that Bangladesh’s economy will perform much better in the next decades than it has in past decades, and there are strong reasons for assuming the opposite.”

Per this World Bank website, it appears that the average per-capita GDP growth in Bangladesh from 1981 to 2010 was 2.6 percent per year.

World Bank

You’re not aware of any “plausible” reasons why the per-capita GDP growth in future decades in Bangladesh can’t be greater than 2.6 percent per year? So there were no “plausible” reasons in Robin Hanson’s paper? And the growth of China and India in recent years doesn’t provide any “plausible” reason?

Are you interested in a bet on your beliefs? For example, starting with 2016, for each year Bangladesh per-capita GDP growth is less than 2.6 percent, as stated on the World Bank site, I’ll give you 10 Euros. For every year it’s more than 2.6 percent, you give me ten Euros. Only the person who has a net loss in money can call off the bet. How about it?

“What the paper doesn’t offer is any evidence at all for believing that the model correctly predicts real world economic processes.”

What model do you think does predict real world economic processes? Maybe the World3/2004 model?

91

Mark Bahner 03.01.16 at 4:44 am

“The World Wildlife Fund says we will need two entire earths to sustain our current lifestyle…”

OK, and the WWF says we currently need 1.6 earths to sustain our current lifestyle. Since we have only one earth, why hasn’t our lifestyle fallen?

92

Plume 03.01.16 at 5:09 am

Mark,

Because the richest 20% of the earth consumes 85% of the earth’s resources, while the bottom 80% consumes just 15%.

Perversely, obscenely, the only thing keeping the developed world going is exploiting and ripping off the developing world — as has been the case for several hundred years now. The only way the math works right now, given the exploitation and extraction by the developed world, is for the developing world to stay off the grid, basically. If it ever consumed as much as we so, it’s game, set and match. Goodbye, humans.

And even as we manage to survive for a bit longer with our current lifestyle, we are permanently destroying our wildlife and fish stocks, ocean ecosystems, the polar icecaps, our mountain glaciers, our waterways, land and air. So we may, for a short time, be able to push off the day of reckoning, with regard to resource depletion, but we won’t be able to stop climate change and overall pollution, which is ongoing and growing as we speak. And that, in turn, accelerates that resource depletion and species extinction, which in turn speeds up climate change and so on. The proverbial feedback loop, and all down hill.

93

TM 03.01.16 at 9:37 am

Mark, sure we can make a bet about Bangladesh’s GDP in 2100 but none of us will be around to claim the prize. Anyway I’m not interested in betting against unlikely scenarios while there are grave reasons for being concerned about the very survival of civilization.

“You’re not aware of any “plausible” reasons why the per-capita GDP growth in future decades in Bangladesh can’t be greater than 2.6 percent per year? So there were no “plausible” reasons in Robin Hanson’s paper? And the growth of China and India in recent years doesn’t provide any “plausible” reason?”

There are no plausible reasons in Hanson’s paper, no. None whatseoever.

Your claim hasn’t been that Bangladesh’s per capita GDP could possibly grow faster than 2.6% per year, your claim has been that growth will roughly double. That extraordinary claim requires arguments, which you haven’t provided. As of China and India, there are presumably reasons for why Bangladesh has not grown as fast (and notice that China’s growth rate is in decline, as are the growth rates of most advanced economies). These factors could possibly change in the future (for the better but also for the worse), but you haven’t made an argument for why this should happen.

In any case, the exchange is getting boring. I have no time for your delusions.

94

Collin Street 03.01.16 at 10:19 am

OK, and the WWF says we currently need 1.6 earths to sustain our current lifestyle. Since we have only one earth, why hasn’t our lifestyle fallen?

Is this really a genuine point of confusion for you?

It’s pretty straightforward: “to sustain” means to continue indefinitely. You can do unsustainable things, you just can’t keep doing them. The statement you’re complaining about is the same in form as “to sustain your current lifestyle, you need 1.6 times your current income”.

Stocks and flows, if you prefer: for something to be sustainable the incomings have to match the outgoings, but if they don’t then — if you’ve got a store of resources, a stock — you can keep doing them. But not “sutainably”: for a little while. Until the store-of-resources runs out.

95

Brett Dunbar 03.01.16 at 11:50 am

More or Less on BBC radio looked at the WWF claim while it added a number of different figures together consumption of most resources was sustainable. The main exception was CO2 emissions which considerably exceeded the Earth’s ability to permanently remove it by laying down carbonate rocks and peat. If the whole world emitted carbon at the north American rate then it would take four times the absorbing ability of the earth to prevent atmospheric concentration from increasing. The data from CO2 completely overwhelmed everything else, so you effectively ended up with a peculiarly stated measure of the excess emission of CO2.

96

Peter T 03.01.16 at 11:55 am

Hanson et al published a paper in Nature which looked at sustainability thresholds. IIRC they calculated that current rates exceed sustainable rates for CO2, nitrogen recycling, marine resources and topsoil depletion, and were close to threshold for a couple of others.

97

magari 03.01.16 at 12:02 pm

Bangladesh is going to be one of the most immediately and hardest hit countries by climate change. Indeed, flooding and salination are already a problem. There’s a very good chance that in 2100 it is more impoverished than today.

98

Plume 03.01.16 at 3:12 pm

Colin @94,

Well done. Very good way to put it.

99

TM 03.01.16 at 3:14 pm

There’s of course the famous Planetary Boundaries paper by Rockström et al. 2009. https://en.wikipedia.org/wiki/Planetary_boundaries

100

Mark Bahner 03.01.16 at 5:21 pm

“Oh, come on. I chose the words “scienzy guys,” as a riff off of the movie, The Martian. The WWF didn’t.”

Yes, I knew you chose the words “Scienzy guys.”

The World Wildlife Fund is predominantly an advocacy organization. (Not that there’s anything wrong with that. ;-)) As Stuart Pimm notes, combining a bunch of numbers for different animals comes up with a easy-to-communicate single number, but it hides the subjectivity behind the number.

For example, look at Figure 8 on page 18 of the report. It shows that the number of species increasing in numbers generally *exceeds* the number of species decreasing in numbers. It’s only through the magic of subjective judgment that they arrive at the conclusion that there has been a 52% decline in the Living Planet Index.

WWF report

101

Mark Bahner 03.02.16 at 3:09 am

“More or Less on BBC radio looked at the WWF claim while it added a number of different figures together consumption of most resources was sustainable. …”

“The data from CO2 completely overwhelmed everything else, so you effectively ended up with a peculiarly stated measure of the excess emission of CO2.”

Yes, so the WWF could simply say, “Atmospheric CO2 levels are going to rise.”

See graphic of “The World’s 20 Largest Ecological Footprints”

102

Mark Bahner 03.02.16 at 3:14 am

“Stocks and flows, if you prefer: for something to be sustainable the incomings have to match the outgoings, but if they don’t then — if you’ve got a store of resources, a stock — you can keep doing them. But not “sustainably”: for a little while. Until the store-of-resources runs out.”

As pointed out by Brett Dunbar in #95. the “footprint” is actually dominated by carbon in the atmosphere. So the index really just means that CO2 in the atmosphere will increase.

103

Mark Bahner 03.02.16 at 5:27 pm

“Your claim hasn’t been that Bangladesh’s per capita GDP could possibly grow faster than 2.6% per year, your claim has been that growth will roughly double.”

No, that’s your assessment of my claim. I wrote in #79:

“I predict the people of Bangladesh in the year 2100 will have a higher per-capita GDP (expressed in year 2016 dollars) than the people of the U.S. today.”

I made no claim as to whether those values represented market exchange rate or purchasing power parity (PPP). Most comparisons between GDPs of countries are done in terms of PPP, since PPP represents a better assessment of the true buying power of individuals in different countries. Per the World Bank, the U.S. per -capita GDP in 2014 was $54,630, and the Bangladesh per-capita GDP (PPP) was $3,123.

Therefore, for the Bangladesh per-capita GDP (expressed as PPP) to exceed $54,630 by the year 2100 would require a growth rate of approximately 3.5% per year.

If you don’t think Bangladesh’s per-capita GDP growth (expressed as PPP) can on average exceed 3.5% per year, I’d be happy to bet otherwise.

Comments on this entry are closed.