The Jevons Paradox: Good News or Bad News?

---

title: "The Jevons Paradox: Good News or Bad News?" date: "2026-09-23" author: "Zhigeng" channel: "frontier" excerpt: "The AI boom has dragged the 'Jevons Paradox,' buried deep in early economic theory, into public view—the same phenomenon, interpreted differently, yields diametrically opposite conclusions. In 1865, Jevons discovered that technological progress raised the efficiency of coal use, yet consumed more coal rather than saving it. Pessimists sound the alarm with it; optimists proclaim the economy's inexhaustible motive force. So is the Jevons Paradox good news, or bad news?" tags: [] readTime: 35

The AI boom—and the anxiety it has stirred—has dragged a concept buried deep in early economic theory, the "Jevons Paradox," into public view. The same phenomenon, interpreted differently, leads to entirely different understandings.

I came to this unsettling problem early, through my research into ecological crisis and sustainable development.

In 1865, the English economist William Stanley Jevons (1835.9.1–1882.8.13) described a phenomenon he himself could scarcely believe: technological progress raised the efficiency of coal use, yet the result was not that less coal was saved, but that more coal was consumed. Later generations named it the "Jevons Paradox."

For more than 160 years, this paradox has been used in diametrically opposite ways, like two faces of a single coin—pessimists use it to sound the alarm; optimists use it to proclaim the economy's inexhaustible motive force.

So, is the Jevons Paradox good news, or bad news?

I. A Discovery That Defies Intuition

In 1865, at the age of 31, Jevons published his book The Coal Question. Coal was then one of only two sources of energy besides plants, the Industrial Revolution's consumption of coal was growing at a furious rate, and British industry faced the danger of stagnation as coal ran out. The book aimed to study the relationship between Britain's coal and its economic development, and its original intention was to explore how to raise the efficiency of coal use and thereby ease the strain on coal resources.

Jevons was a rigorous economist, and a man acutely sensitive to numbers. Following the habitual line of thought—"raising efficiency saves coal"—he investigated and reasoned his way to a conclusion that he himself found hard to accept.

His finding was this: saving coal increased profits and lowered product prices, attracting more investment—so the total amount of coal used, far from decreasing, actually increased.

Jevons confessed in the book: "When I began this inquiry, I had little idea of some of the results"—"to suppose that an economical use of fuel is equivalent to diminished consumption is a confusion of ideas, and the very contrary is the truth," and "it is wholly a confusion of ideas to suppose that the economical use of fuel is equivalent to a diminished consumption. The very contrary is the truth... It is the very economy of its use which leads to its extensive consumption. Nor is it difficult to see how this paradox arises."

He even said that he would be very glad if his conclusion turned out to be wrong.

A scholar who most hopes his own conclusion is wrong—this shows he has clearly seen the disquieting thing his conclusion implies.

This "disquieting thing" can be unspooled into a clear chain:

Efficiency rises → costs fall and profits grow → demand expands, investment expands → total demand, far from falling, rises.

Some scholars call this the "rebound effect": what you save gets used back on a larger scale.

The terror of this conclusion lies in the fact that coal is finite. The more you use and the faster you dig, the sooner it runs out. At that time, oil had not yet entered the industrial energy picture, and the exhaustion of coal meant the decline of industry.

In the 1960s, the age of oil began. With energy thus carried forward, industry flourished all the more because it had shed its dependence on coal. This is precisely the optimists' basis—technological progress can solve every problem: when coal runs out there is oil, when oil runs out there is natural gas... What the Jevons Paradox portends is not economic tragedy at all, but the economy's inexhaustible motive force.

Here the optimists overlook one problem: not only is coal finite, not only is oil finite—the Earth itself is finite; and not only are resources finite, the Earth's capacity to absorb garbage and waste heat is finite too.

II. The Six Orders: The Paradox's Logical Extension

Jevons himself completed the paradox's two most basic orders of expression, later economists developed it into three orders, and in my essay "Self-Dissolution: The Extreme Feature and Ultimate Dilemma of the Risk Society," published in the fourth issue of Engineering Studies in 2020, I further organized the paradox into a six-order formulation. Here I set it out briefly.

The First-Order Rebound

Raising the efficiency of a given energy source (coal, for example) brings growth in profit, and the increased income is reinvested in expanded reproduction; the rise in efficiency also lowers the price of products (pig iron, for example), thereby enlarging sales volume and further increasing profit. All of these directly increase the total consumption of the energy source whose efficiency has been raised.

The Second-Order Rebound

Improving technology and equipment in order to raise energy efficiency enlarges the demand for equipment, and even invents new equipment (improving the steam engine, for example, brought the steam plow and the steamship), indirectly increasing energy consumption.

The Third-Order Rebound

Rising efficiency increases income, raising producers' capacity to consume other goods and even to invest in other fields, stimulating the consumption and expanded reproduction of related goods, so that energy consumption—and the consumption of other resources bound up with it—extends into ever broader domains. The increment of consumption extends from one resource to many resources, and finally to all resources—including the demand for such broad resources as waste-absorbing capacity and ecological equilibrium.

These three orders are known in many studies, in sequence, as the "direct rebound," the "indirect rebound," and the "comprehensive rebound."

The fourth through sixth orders are my extension, expressing "the way the results of human action betray or even dissolve its motives."

The Fourth Order: The Manifest Aggressive Crisis

Certain technological, engineering, or socioeconomic projects with clear application-oriented goals—curing the sick and saving lives, for example: the better they are done, the faster the population grows, in what might be called "the better, the more disastrous." Reducing pain and extending life is a kind and beautiful pursuit, yet overly rapid population growth is at the same time an acknowledged human crisis; and the series of problems the population crisis brings, to a large extent, offsets the very welfare that reducing pain and extending life confers.

The Fifth Order: The Hidden Aggressive Crisis

For example, "ruin-bringing knowledge" and the indiscriminately growing scientific frontier—at the time of the research, it may aim only to reveal nature's mysteries or to seek out a technical path, without any clear applied expectation, and may even disregard application altogether, yet it conceals a ruinous potential and makes possible future ruinous applications.

The Sixth Order: Human Culture Itself Bears Within It a Self-Destructive Tendency

At this point, the Jevons Paradox has already leapt from a phenomenon of resource economics into a question about human civilization itself. (The last three orders and the "aggressive crisis" will be addressed in a separate essay.)

III. How the Optimists Turn the Paradox into a Reassuring Pill

On the question of employment, the Jevons Paradox is used, repeatedly and with absolute conviction, as a piece of good news.

This usage rests on a layer upon layer of argumentation that seems, on its face, unassailable.

Its theoretical backing comes from an authority in economics. In his Principles of Economics, the world's best-selling textbook, Mankiw takes up what he calls the "lump-of-labor fallacy": people mistakenly believe that the total amount of labor society needs is fixed, and that every task a machine completes permanently subtracts a piece of work left for human beings. Mankiw argues that technological progress eliminates specific jobs while simultaneously lowering costs, raising efficiency, and creating new demand and new industries, so that total employment does not decline—workers merely move from old jobs to new ones. His classic examples include: after agricultural mechanization, the American agricultural population fell from an overwhelming majority to roughly 2%, yet no permanent mass unemployment resulted; ATMs did not eliminate bank tellers—on the contrary, because the cost of opening new branches fell, banks hired more tellers.

Directly naming the Jevons Paradox is the technology commentator Benedict Evans. In July 2023, in his essay "AI and the automation of work," he wrote: "We've actually been applying the Jevons Paradox to white-collar work for 150 years." He traces this history: the typewriter let one clerk produce more than ten times the text, yet it did not reduce the number of clerks—it hired more; the adding machine let one person do the arithmetic of five, yet there were more bookkeepers; after the spreadsheet (Excel) appeared, employment in accounting rose rather than fell. His conclusion: "Automation plus the Jevons Paradox means more jobs."

Empirical support comes from major institutions. Reports from the McKinsey Global Institute repeatedly stress that less than 5% of occupations can be fully automated, and that in previous technological revolutions, the loss of agricultural population and the declining share of manufacturing employment were both accompanied by the emergence of large numbers of new jobs. The Asian Development Bank's annual report, titled "Will robots take our jobs or create them?", reaches the conclusion that one should "remain optimistic about the prospects for job creation."

And just recently, the Jevons Paradox has been pushed into a new and, indeed, the hottest field of all—AI computing power. In its latest Asia-Pacific AI technology research report, UBS writes in black and white: based on the "Jevons Paradox," the overall demand for AI computing resources in the future will continue to expand as the cost of inference-side computing power declines. The report restates the paradox's definition in full—frontier technological progress raises the efficiency of resource use and lowers unit costs, thereby stimulating a large-scale expansion of market demand, ultimately causing the total consumption of the resource to rise rather than fall.

Alibaba CEO Wu Yongming's remark at the 2026 Apsara Conference—"In the Industrial Revolution, machine power ultimately took on 99.9% of human physical labor, and in the future machines will take on 99.9% of thinking work," and "leave the drudgery to machines, leave creativity to humans"—completes a full "good news" narrative.

Laying out this argumentation, one sees that it runs from textbook to blog, from institutional research report to entrepreneur's speech, across every level of authority, with a highly consistent message and a highly isomorphic logic:

Efficiency rises → demand expands → the total does not fall but rises.

This sentence is the very mechanism of the Jevons Paradox. What the optimists do is take a mechanism that originally belonged to "energy" and transplant it, unchanged, onto "employment," then announce: look, technology will not eliminate jobs; it will only create more.

It sounds very reasonable—but a premise has been overlooked.

IV. What Have the Optimists Skipped?

First, the optimists fix their eyes on job opportunities while forgetting the physical boundaries of the underlying resources that support those opportunities—beneath computing power lies the chip, beneath the chip lie materials and energy, and beneath materials and energy lies the boundary of the Earth.

If these boundaries are breached, to what can work attach itself?

Some may say: no problem—when Earth's resources run out we can mine space, and space's resources are inexhaustible.

At this point I must remind you once more: not only are Earth's materials and energy finite; the broad resources for absorbing garbage and waste heat are finite as well, and the excessive accumulation of garbage and waste heat since the Industrial Revolution is now shaking the foundations of human survival.

The ecological equilibrium of the Earth's biosphere is a precious result of billions of years of evolution and brutal self-adjustment by the solar system and the Earth. Recklessly altering individual indicators within it will cause the ecosystem to disintegrate and collapse.

This is also why I argue that humanity should expand its living space, go beyond the Earth to make its attempts, use space resources in place, and not bring them back to Earth for use.

Second, will job opportunities alone really prove so optimistic?

The optimists take only the surface of the Jevons Paradox—"efficiency rises, demand expands, the total rises"—while quietly discarding its foundation, that is, the premise that makes the paradox a paradox: finiteness.

In the field of energy, this "finiteness" is concrete and cruel—coal can be dug out, oil can be burned up, and so "the higher the efficiency, the greater the consumption" is a disaster. The Jevons Paradox is unsettling precisely because it proves that, in a world of finite resources, efficiency itself may accelerate the exhaustion of resources.

So when the optimists transplant this mechanism onto employment, what is the corresponding "finite resource"?

The answer: human time, human attention, and labor itself.

The optimists generously assume that, after efficiency rises, society's demand for "things for humans to do" will be, like the nineteenth-century demand for coal, infinitely elastic and boundless in space.

In the previous industrial revolutions, the Jevons Paradox worked because what was replaced was always "one part of a person's capacity"—

The typewriter replaced the act of "copying," so the displaced person was promoted from "copyist" to "clerk using a typewriter," doing more and more complex things;

The adding machine and the spreadsheet replaced the act of "manual calculation," so the displaced person was promoted from "abacus operator" to "financial analyst," still doing more and more complex things.

But AI is different. What AI replaces is thinking, judgment, and creation. When a machine can not only copy but also write; not only do accounts but also make decisions; not only write code but also architect systems.

What AI replaces is labor itself, and even the very existence of human beings.

This is the watershed between "good news" and "bad news."

The optimists have always assumed that "the Jevons Paradox will always hold," yet they have never answered a more fundamental question: when labor itself can be produced limitlessly, cheaply, and rapidly, will the chain of "efficiency rises → demand expands → new jobs" still, as it has for the past two hundred years, automatically and endlessly continue?

Reasoned through, the answer to this question also has two faces:

Following the optimists' line, the Jevons Paradox will, for the first time in history, lift humanity onto a higher rung: ascending beyond humanity itself—humanity evolving into the cyborg—which means enormous technological risk and the screening of survival opportunities, and cannot escape the risk of total destruction.

Following the pessimists' line, what the Jevons Paradox portends is that a frenziedly developing AI sucks the Earth's resources dry, destroys the ecosystem, and marches toward total destruction.

Jevons once said that he would be very glad if his conclusion were wrong. Today, I dismantle this paradox against the background of the new AI era, and the result is that whether one is glad or grieving, one ends up grieving. I too hope my deduction is wrong.

There is no perfect ending, only a faint glimmer of rebirth through the flames—and that glimmer can by no means light the way for everyone.

The narrow passage of civilization's "great funnel" lies here; the answer to the Fermi Paradox lies here.

Have you, dear readers, ever thought that the Jevons Paradox is the answer to the Fermi Paradox?

In the next issue, I may dismantle an even grimmer riddle. Stay tuned.