Hacker News Reader

> So economists do the same thing, they take some ratio of numbers that seem to be tracking the 'size' of the economy and then run that trend out 50, 100, 500 years.

Very few academic economists do 30+ year extrapolations, let alone 50, 100, 500. The vast majority of long-term growth macro-economic literature are attempts to explain/model the large observed differences in outcomes in GDP/capita between different countries, and hopefully distill some useful policy advise from that. Forecasting beyond even 5 years or so is a very niche activity in academic economics. (Source: PhD economics and former model builder for pension fund and souvereign wealth funds)

Fair point, I withdraw any assertion that in published papers economists over extend their numbers. It has been my experience that in the reporting on those papers I have seen much less restraint. And yes, I'm aware that survivor bias may mislead in this regard.

Buckminster Fuller made a similar observation or prediction even earlier, in the late 1930s. He called it ephemeralization, https://en.wikipedia.org/wiki/Ephemeralization, the idea that increasing technological efficiency will allow us to do evermore with less.

A possible counter-acting force is Jevon's paradox (a generalized economic form of Wirth's Law) but the linked empirical result suggests ephemeralization is the stronger force. Although, it is likely true we can do much more to counteract Jevon's Paradox like phenomena.

The Jevons Paradox [1] is named a "paradox" as it is an unusual, apparently contradictory outcome. It can be considered a special case of the rebound effect [2] where the rebound is greater than 100%. Unlike rebound effects in general, a Jevons Paradox is not a routine outcome of any new technological efficiency gain. It is something that needs to be observed from data rather than assumed.

In my experience, the Jevons Paradox seems to be more widely brandished than understood in internet debates about energy and resources [3]. If Alice claims e.g. that American electricity production is getting more efficient, Bob may "refute" her point by citing the Jevons Paradox without mustering evidence that the Jevons Paradox applies to recent trends in American electricity consumption.

[1] https://en.wikipedia.org/wiki/Jevons_paradox

[2] https://en.wikipedia.org/wiki/Rebound_effect_(conservation)

[3] Your comment did not misuse the Jevons Paradox. You noted that ephemeralization seems to be winning.

There is a vast pent up global demand for electricity around the world. Africa needs electricity, China needs electricity, India, etc.

The world is changing but at the moment that electricity largely comes from burning fuel. And if it makes sense to mine fuel when the payoff is X, then if still makes sense when the payoff is X + something (as it is when efficiency improves).

It is a very safe assumption that Jevons's paradox applies to electricity consumption, there is no need to show it. Take a look at US coal production [0]. There is no evidence there that energy use from coal will decrease with improved technology. That is 150 years of technological innovation and we're basically seeing a linear uptrend in use. If domestic use levels off there will be exporting. That production is not going to go down without an uneconomic political intervention or literally running out.

Now I'll agree that Jevons's paradox does not apply to everything [1]. But energy in general and electricity in particular it is painfully obvious that there are more potential uses than there are efficiencies in the world. Give me cheap enough electricity and I can do anything; its use will not be suppressed by reducing need for it in specific cases.

[0] https://en.wikipedia.org/wiki/Coal_mining_in_the_United_Stat...

[1] Also my name isn't Bob.

The decline of US coal production is right there in the page that you linked to:

https://en.wikipedia.org/wiki/Coal_mining_in_the_United_Stat...

American coal production peaked in 2008 at 1.172 billion short tons. Domestic use peaked in 2006 at 1.150 billion short tons. As of 2018 those numbers are down to 0.755 and 0.678 billion tons short tons respectively.

The EIA's latest Short Term Energy Outlook predicts that total US coal production for 2019 will be 0.679 billion short tons and for 2020 will be 0.603 billion short tons, 48% below the peak.

If you look at the "components of annual change" graphic in the EIA report, you can see that coking coal (used for making steel) is up slightly from 2017-2020. It's coal consumption for electric power that is in persistent decline.

https://www.eia.gov/outlooks/steo/report/coal.php

Coal supplies about 25% of American electricity in 2019 and is expected to decline to 22% in 2020:

https://www.eia.gov/outlooks/steo/report/electricity.php

Percentage-wise, coal reached its all time high in the mid-1990s, at over 50% of American electricity generation:

https://www.mjbradley.com/sites/default/files/MJBAcoalretire...

The Jevons Paradox observes the introduction of more efficient technology in an existing system but finds a "paradoxical" uptick in input consumption. The introduction of more efficient coal burning machinery in England was followed by higher coal consumption. That surprised Jevons.

The broadening of energy use (e.g. India developing like England did) certainly also increases demand for energy inputs, but it's a different case than the Jevons Paradox.

Which is meaningless unless you believe that energy efficiency has only happened in the 2008-2018 period, which is not true. I happened to be working in the coal industry for most of that time and the decline isn't because people are using 40% less energy. It has nothing to do with increased efficiency. It has to do with:

1) Political action.

2) Alternative energy sources getting more competitive. Particularly natural gas, renewables and US shale oil.

You're arguing that use-of-energy efficiency can cause a decline in coal production. That is wrong; if you use energy more efficiently I'll mine more coal - I've met the people who make that sort of decision. Coal is declining for other reasons - mainly that alternative sources are getting cheaper (which is not using energy more efficiently).

I did not say that greater efficiency in coal use was the cause of US coal's decline. My last post was contra to:

Take a look at US coal production [0]. There is no evidence there that energy use from coal will decrease with improved technology.

There is ample evidence that improved technology is driving declines in American coal use/production. The combined cycle gas turbine made it possible to extract more electricity from a megajoule (thermal) of natural gas than a megajoule of coal. Hydraulic fracturing made natural gas cheap in the US. Improved technologies and economies of scale made wind and solar electricity more efficient and cheaper. These are all electricity production technologies. They are what what I had in mind with my up-thread Alice and Bob example. US electricity production is getting more efficient (more electricity per ton of material inputs consumed); Bob believes that any efficiency increase must be more-than-offset by increased electricity consumption, and exclaims "Jevons Paradox!" before even reviewing the numbers.

I was just going to mention Ausubel. He gave a great talk for the Long Now a few years back. Video here: http://longnow.org/seminars/02015/jan/13/nature-rebounding-l...

> The systems guy in me has experience that all exponential trends are s-curves

A friend of mine visited a financial planner inquiring about saving for his children's university education.

The planner took the last the last 20 years of education inflation and extrapolated it out 15 years, and recommended my friend put away 2.5k per month today per child today to cover it.

My friend didn't hire the planner, on the reason that we will probably be post-revolution (which would cause this curve to become an S) before we get to the point where you need to save 30k per kid per year over 15 years to pay for university.

Did they consider that the USA send all their production to China due to cheap labor and few environmental regulations? Now they are consuming more but using less raw materials. Compare the number of toys that a child has today to one 30 years ago.

> all exponential trends are s-curves

What does this mean for the stock market? And debt? Or really, what does an economy look like when it's done growing, but stable?

Look at the Japanese economy for an example of stability without growth.

> Its nice to see this reality (of how extrapolating is bad because things change)

...but when they do that same thing on earth temperature everyone loses his mind

"Gee, why is there a pool of water on the floor ever since the rain started? Do you think maybe the roof is leaking?"

"Don't be silly! Just because there is water dripping from the ceiling doesn't mean there is a leak in the roof. Things just get wet sometimes. I'm sure there's a wonderful explanation besides the roof leaking. No sense in hiring someone to come look at the roof either. That would be a waste of resources."

case in point - did I say I am not believing that climate is changing? but no, as soon as one contest the wild exponential predictions, he gets scourged.

The pessimistic outcome of climate change isn't "wild exponential predictions", though. There are well-understood mechanisms, the margin for temperature increase beyond which modern society collapses is pretty thin, the exponent comes primarily[0] from our economy which has an assumption of exponential growth built-in and will collapse if it suddenly turns into s-curve, and we also have a model of a runaway warming feedback loop as our neighbor in the Solar System.

Also worth remembering is that climate change isn't really about climate. It's about our society. Earth doesn't care about temperature, and life on Earth will happily survive whatever we can possibly throw at it this century. All that matters is that climate change puts pressure on societies around the world, and with enough pressure our entire civilization will collapse, and you and me and everyone else we know will die a horrible death.

--

[0] - There are also environmental feedback loops ready to take over the growth of average temperature that have gained attention recently.

What climate change effect [0] do you think will cause societal collapse or result in mass horrible deaths?

[0]: https://climate.nasa.gov/effects/

> For example, at the end of the last ice age, when the Northeast United States was covered by more than 3,000 feet of ice, average temperatures were only 5 to 9 degrees [Fahrenheit] cooler than today.

If 4 degrees Celsius of global cooling (relative to the current climate) can cause large parts of the US to be covered by 3,000 feet of ice, then it's not unreasonable to think that 4 degrees Celsius of global warming could have an equally destructive effect on agriculture and society.

Hundreds of millions of people migrating from areas that become uninhabitable from sea level rise and increased temperatures, trying to get into any country that's less affected. Europe just almost tore itself apart over few million war refugees, imagine 100 times more people, with nothing to lose and a valid argument that it's all our fault they don't have a home now.

Food pressure, as arable land is lost to the sea and temperatures[0], plus collapsing ecosystems as insects and plants with high vulnerability to temperatures die off. Food prices increase, risk of food shortage increases too, and so does social unrest.

As some land is lost, other land becomes available, "defrosted" by the changing climate. With reduced agricultural capability and migratory pressures, nations will want to compete for both that new land and resources available on it. This could lead to war.

I could elaborate, but let me instead tell you to look at the problem in reverse: think of how little it takes to cause "mass horrible deaths". Most people live in cities now, and most cities are fed "just-in-time", which means there's 3-5 days worth of food in them. Imagine what happens if you break that supply chain for more than a week. Imagine this happening to multiple cities in a country simultaneously.

--

[0] - You can't e.g. think that a +4°C Poland can just switch from wheat to bananas. Bananas may like the new summers, but won't survive the winters.

The heat-trapping behaviour of CO_2 was first accurately measured in the 19th century, and all climate change models do actually predict convergence, not linear rise. It's just that the point of convergence is not one compatible with civilisation as we know it.

Exactly. If global warming said the typical outside temperature in Seattle doubles to almost 600K then it'd be fair to argue that's crazy people extrapolating - water isn't even liquid at that temperature, no model we have could make sense. But actually the model says it will go up say 2 or 3K. Much less like a wild extrapolation, and unfortunately still completely terrifying because that is already far too much.

2 or 3K really isn't enough for Seattle.

How does anyone know what temperature rise is or isn't compatible with civilization as we know it, within the range of predicted temperature rises? One of the problems is that civilization is dynamic. New technologies come along, people change their habits, etc.

The effects of temperature rises on crop yields could lead to very negative outcomes, given expected population growth and the hard limit of humans' minimum caloric intake.

https://www.scientificamerican.com/article/for-crop-harvests...

Perhaps rich countries can mitigate the likely increase in food prices, but it could destabilise large regions of the globe.