Author Options:

Jevons Paradox Answered

I recently read a thought-provoking article in the New Yorker: "The Efficiency Dilemma" by David Owen. The main gist of the article is that as our machines use less energy, we use them more, thus negating any environmental benefit we hope to achieve. This is known as "Jevons Paradox". You can read the abstract and article here, but you have to be a subscriber for the full article. You can read a commentary, with excerpts, here.

Please discuss...



7 years ago

The "Paradox" highlights some important issues that could be naively overlooked, confusing and leading to incorrect assumptions such as "energy conservation is futile"wiki.
  1. Efficiency in this sense is economic efficency, (~ usefulness / cost).  This is not cleaner or less-polluting efficiency.  Most environmental and societal impacts are not captured in "cost", for better or worse.
    • More "efficient" is not necessarily better for the environment, that is confusing the issue.
    • Additionally, "efficient" does not mean less-quantity for same output (that's only one possibility).  Simply dropping taxes on gasoline would make the cost less, thus encourage more use, resulting in Jevons paradox.
  2. This is a paradox only in the sense of a limited field of view of supply and demand.  The S-D curves of all substitute goods need to be considered for a correct analysis of impact.  More efficient coal will result in more coal purchase, at the same time reducing the purchase of oil.  (Does that help the environment?)
  3. Finally, the paradox implicitly assumes scarcity.  This is fairly correct, but in the mathematical limit, could be wrong.   If the amount needed to perform work (physical efficiency) was << unit of cost, the good would no longer be scarce and consumers would have more than they need, essentially no longer consuming more at better prices.
The paradox is useful to highlight / avoid naive conclusions. 

Making gasoline engines more physically-efficient, thus more $-efficient, is not, perhaps, the best policy to lower combustion emissions.  A better RD investment would be increasing the physical-efficiency of solar powered engines, and thus the $-efficiency.  This would encourage more solar-energy use (per Jevons Paradox) which would create less combustion emissions.  A more realistic economic policy for reducing combustion emissions.

The Wiki article also mentions

However, ... in order to increase energy conservation, fuel efficiency gains must be paired with some government intervention that reduces demand (e.g., cap and trade, fuel tax or carbon tax).

Which I suggest is also naive.  Again, rather invest in substitute goods that have the environmental benefits you want.  Taxing away savings is not a good way to encourage development (but does expand the budget...)

Excellent analysis and summary! Thank you very much, and thank you especially for pointing out what I had overlooked, that "efficiency" is used in the economic sense, not the physics sense.

De nada.

It's interesting how economics is rather sloppy with definitions, particularly compared to physics or mathematics.

I lived in NYC during the boom of quants and I always feel they are a great example of how business moves faster than precision. The field suffers, from that point of view.

Not acceptable situation in science, but of course the problem there is the budget constraints!

I guess the worst of both would be Government...

ok, here's my 2 cents -

it sounds to me (also not a scientist, so please be gentle with me) that the article is pointing out the possibility that increased efficiency is counterproductive. and i can understand the idea of consumption going up when costs go down. but i feel like "increased efficiency" is being painted as the bad guy and i have to jump to his defense and point an accusatory finger at the religion of unlimited growth.

my hybrid car does not get any more miles than its predecessor and i do not do more laundry than i used to because i have an HE washer and dryer. i totally recognize the larger scale version that results in "more population, developing countries becoming more ubanized and industrialized, and so on" but i think that's because the idea of unlimited growth is treated like an untouchable religion that's protected from any and all criticism. it is not ok to waste food, just because you can afford to buy more. it is not ok to run the air conditioner all summer long because you can afford it - even if you live in the desert. in fact, i'll even argue that it's not ok to live in the desert. but i've been told a number of times that it's not ok to express these opinions, because people should be able to do what ever they want. if they have money to make an otherwise uninhabitable place inhabitable (and then build a golf course there!), i have no right to stop them - regardless of how wasteful i think it is. i have no right to judge families with children in the double digits. and i must be a communist if i think it's ok to talk about population control. [shrug]

just saying - i think that if we could paint "unlimited growth" as the bad guy, we could let "increased efficiency" off the hook.

You are exactly right! See my quantitative description much farther down. Jevons' paradox basically boils down to the observation that, using my own notation, N*(1-E) is an increasing function of time.

Even if the "efficiency" E (whatever you're measuring as an improvement) gets closer to one, so that the "inefficiency" (environment cost or whatever) goes down, the "unlimited growth" pressure of society increases N (the amount used, measured however is appropriate for the problem) to more than compensate.

I think you're exactly right that "increased efficiency" is not, and should not be, considered the villain. Decreasing the per-unit cost to the environment, or to society, is always useful and productive. Even displacing those costs, say from population centers or agricultural land, out to desers, or the abyssal ocean, or space, is still beneficial to society.

What Jervons' observed, and what seems to be a general aspect of human societies, is that we still manage, over time, to take advantage of improvements, rather than having those improvements stabilize.

"What Jervons' observed, and what seems to be a general aspect of human societies, is that we still manage, over time, to take advantage of improvements, rather than having those improvements stabilize."

that sounds much more accurate than my assumption (about efficiency being vilified). thanks! :)

You may not be wrong about the tone of the New Yorker article, however. I've noticed a number of authors being fairly negative about new, "environmentally friendly" technologies, and using the argument of Jevons' paradox as their basis. I believe that is short-sighted, and more of an argument to maintain the status quo (with it's concomittant effects due to growth), than to attempt to improve matters.

oh, ok! so i wasn't just being hyper-sensitive :) also good to know.

so how do we redirect human nature from over-consumption?

hehe i call that revelation a "duh" event.

it sort of follows the same principle of hard drive storage (i can't remember the "law" name off hand)...as hard drive storage space increases, our ability to fill them does as well.

i would imagine though, that at some point the efficiency gains would negate this principle eventually

well, i'm not a scientist (like yourself) so i doubt that i'll be reading all of those.

so, by your statement, there will never be a time when there is any device that has a neutral effect on the environment?

No. By my statement, there will never be a time when the efficiency of technology will be such that the aggegate use of that technology by human won't be as detrimental, or more, than the effects of older, less efficient techology.

The point of Jevons' Paradox is not that the technology itself is more detrimental when it is more efficient. The point is that when technology is more efficient (or less expensive), then society will use it even more than it was using previously. Here's the quantitative statement:

Let E be the "efficiency" (some number between 0 and 1); then the "inefficiency" (e.g., waste products, CO2 emissions, heat, whatever you want) is 1-E; as efficiency improves, inefficiency decreases. Let N be the "aggregate usage" (some number related to population, or number of units, or hours used, or whatever).

Then Jevons' Paradox says that for any Enew > Eold,
you will always have Nnew > Nold, such that

Nnew(1-Enew) > Nold(1-Eold)

As efficiency improves, we use more, not less. Some of this is due to increased population, some is due to increased production and productivity, some is due to the integral of individual choices made (my car gets better MPG, so I can drive farther than before).

well, let's go completely off the rails here and come up with some hypotheticals that may or may not be completely scientific or 100% accurate. but this is just so i can understand the concept.

let's assume that someone semi-perfects the hydrogen fuel cell (the type where you put in hydrogen and it spits out water) and it becomes standard for all vehicles. at some point (assuming that people just stop looking for something better) the majority of vehicles would end up using this as their mode of propulsion. it's my understanding that this part of the process would reduce the environmental impact of the vehicle in question (as far as direct environmental impact from the vehicles power production itself) by a pretty large margin over current ICE tech.

i guess that would leave the 1-E to come from the production of hydrogen (which, it's my understanding is the biggest stumbling block on the road to hydrogen cars....besides peoples concerns about exploding), and the production of the equipment in and of itself correct?

based on the fact that the emissions of the vehicle in question would be water (water vapor), you wouldn't see a gain in environmental impact from increased usage (at least not from the vehicle itself) right?

at that point wouldn't you have basically cut out one of the contributing factors over current ICE usage? since with ICE you have to compute for the manufacture of the car, the manufacture of the fuel, and the emissions of the car?

or am i still WAY the hell off?

Excellent work. Your analysis is correct, as far as it goes. By switching completely from fossil fuels to hydrogen (which could either be fuel-cell or IC), you eliminate (E -> 1) the direct environmental impact of burning fossil fuels in vehicles.

However, that analysis is incomplete. It turns out that the major source of hydrogen gas is not electrolysis of water, but in fact the chemical decomposition of...wait for it....fossil fuels (there's another topic around here where Goodhart posted a link). So you haven't actually eliminated that impact, just moved it from the end user to farther up the supply chain.

The same unpleasant side effect, by the way, applies to either electrolysing water for hydrogen, or for charging battery-storage electric vehicles. Our current electrical infrastructure uses primarily fossil-fuel plants, because they can be better controlled to provide constant output as well as to respond to rapid changes in demand.

Thus, at least some of the direct impact of ICEs is not eliminated by changing vehicle technology, just centralized at power distrubution centers. This is still a good thing -- emissions pollution tends to stay localized: instead of millions of cars pumping pollution into large population centers, you have hundreds of power plants spread over a large area, pumping pollution into low-population areas.

well, then if we move the efficiency up the chain and find a better way of getting the hydrogen (i've got no clue what that is), that's more efficient than the current best method (i.e. decomposition of fossil fuels) that has (at the very least) a lower 1-E at the source than the current method? then whats?

Then you'll see more total usage (more population, developing countries becoming more ubanized and industrialized, and so on). The point of Jevons' paradox is that even as 1-E decreases, N increases (over time) to not just compensate, but to exceed the previous equilibrium.

i dunno, guess i just can't wrap my brain around infinite itterations

so, at no point during the chain of advancement would there possibly be a way to achive lower environmental impact, no matter what you do, ever, for any reason.

slightly depressing at best.

i would have to imagine that the root of my original post would have to follow through eventually though, given current and forseeable future technologies the equation is obviously applicable. but isn't one of the main points of advancement in power production to achive something that's as close to ecologically neutral as possible?

say, a cold fusion reactor on the moon that's big enough to power all of the electrolosys facilities on the planet for the next 1,000 years, and a really long extension cord?

I think the difference is between immediate and longer-term results. Suppose that we could, today, magically replace all ICEs with fuel cells. Then, yes, we would have a substantially lower environmental impact, just as your intuition suggests.

What Jervons observed about coal, and which is a more general result, is that once you have that new technology in place, then over time people use it more and more, and do other things with it, which offsets the initial improvement.

Another good example is the introduction of electricity to a rural economy. At first, there's a great gain both in production efficiency (an electric pump can collect much more water than people using hand-cranked wells, for example). But then people start staying awake after sunset, using more electricity, and needing heating and whatnot. They start using all the water they can pump out, draining aquifers, and so on.

The relation is not instantaneous, but delays, sometimes by years, sometimes by decades or more. Does that help?

alright, so, let's go with CFLs.

while most people have adopted their use, i would imagine that most people don't use them more (i.e. have their lights on longer just because they can), and i would postulate that most people don't use more of them (install more fixtures because they can). so, they're using less juice, and i don't remember anything specific stating that they're more environmentally damaging to make.

also, how does the model account for an "if it was still the old way but everyone is still using the same as the new" in it's calculations? like, if we reversed the CFL scenario, and everyone who is now using CFLs (lets say for the sake of example that all lightbulbs are CFLs right now) went back to incandescent lights today and used them the same ammount as they do the CFLs? wouldn't that be using more juice than the CFLs?

With CFLs, the "impact" is shifted to another problem. Adding extra mercury compounds to the environment. The impact of greater use may be an initial savings of energy used, but in making them and then eventually disposing of them, the impact returns.

IF I understand this topic properly :-)

right, but is the increased mercury after disposal greater than the electrical savings?

I suppose, that depends on how much energy it takes to make them, and how many people / animals die of the poisoning,

right, that's what i'm trying to wrap my brain around....is this thing as hard/fast as thermodynamics/conservation of energy, or is it theoretically possible that the equation won't apply to something eventually

It is pretty much based on conservation of energy though, isn't it? I mean: GIGO?

is it though? i mean, i don't see where ecological damage is part of the conservation of energy equation. since (if i'm understanding as well as i think i am now) the whole proposition is about the fact that you get no gain in ecological benefit (or you achieve losses) by making things more efficient.

I meant that effort in = effort out - efficiency losses. The mercury would add to the efficiency losses if it lowered the population enough ;-)

i guess that's what i'm getting at though, does that negate the efficiency gains?

It depends on the manufacturing process again. I am not familiar with how difficult it is to make them.

But yeah, if my light bulb caused me to die, I might consider that a loss in efficiency :-)

You don't have to be a scientist to think like one, nor do you have to be a scientist to read actual research, rather than some bowdlerized summary or political interpretation of that research.

no but it helps to have a scientific background when reading scientific papers....especially if one doesn't want to fall asleep whilst reading.

That's true, but not essential. Being generally well read, and having an ability to use a good (unabridged or OED-level) dictionary, or to look up technical terms online, are sufficient. Understanding how end notes and citations work, and using them to follow a trail of research results, is also a useful skill.

If you aren't a specialist in a given field, it is better to look for review articles, which provide a summary of the current state of a given topic, rather than individual articles on research in that topic. The latter are more likely to be short, narrowly detailed, and to assume a lot of prior knowledge (they're written for other researchers, not for general readers).

alright, i'll just say it. "i'm not a scientist" meant "i'm too damned lazy to read most of your sources, but i read one and got the gist" :P

Excellent links, as usual.

Google Scholar is a great resource. My only (major) complaint is that too many of the links are to pay-per-view sites like Elsevier, so users cannot actually get to the primary research :-( At least it gives you the citation references so you can go to a good local college library and read the physical papers.

I have a good start on my holiday reading list now...


7 years ago

I thought that this article would spawn some lively and perhaps vitriolic discussion within the Instructables community. I certainly found it thought-provoking.