Is the Fast Energy Transition Falsifiable?

“No amount of experimentation can ever prove me right: a single experiment can prove me wrong”. Albert Einstein

It seems easier to falsify the belief that the energy world is not changing, than it is to falsify the claim that it is.


The energy switch, from fossil fuels to manufactured, non-carbon, wind/solar/battery conversion technologies will be far quicker than previous energy conversions due to compelling technology and economics, intensified by social license issues driven by climate change, air quality and energy security concerns. The empirical data that supports this theory –  comprising learning curve cost improvements, investor reaction to peaking fossil fuel demand and social license and climate change momentum – form a unifying explanation to support the theory that this energy transition will be fast and disruptive.


Any worthwhile theory needs to be falsifiable if it is to be taken seriously. The Fast Energy Transition theory is predictive and clear enough to be falsifiable – many tests and observations would undermine the theory if they were to be validated. So the theory, whilst still to be fully verified, is robust enough to handle falsification challenges.

In fact  it seems easier to falsify the belief that the energy world is not changing, than it is to falsify the claim that it is. Models of business as usual (BAU) are being constantly falsified by the objective data of wind, solar and EV uptake happening “faster than expected”. This is not merely incumbents getting a few BAU forecasts or scenarios wrong; it means the whole BAU or gradual transition hypothesis is constantly being falsified, and it should be rejected, in favour of the fast transition which has withstood these tests. 

Prologue: Popper and Falsifiability

A theory is a tested, well-substantiated unifying explanation for a set of observed or observable phenomena.

A hypothesis, on the other hand, is a proposed explanation which has far less empirical, observable evidence (so far) to support it.

Karl Popper proposed the notion of falsifiability as a means to advance hypotheses and conjectures into more established scientific theories.

That is, any good theory should make clear, testable predictions that can be endorsed or falsified via accessible methods.

Popper further proposed that verification and falsification had an inherent asymmetry, because universal statements can’t be derived from singular statements, but can be contradicted by one.

As Einstein summarized it: “no amount of experimentation can ever prove me right: a single experiment can prove me wrong”.

Importantly, falsifiability does not necessarily abolish or terminate the existing theory: it may just amend or extend it to a better one that still holds much of the detail of the prior theory.

For example, the theory of light moved from ancient Greek and Newton’s notions of light particles, to be replaced by Young’s theory of waves in an ether, which was then disproved via the Michelson-Morley experiment, and then further upgraded by Maxwell’s electro-dynamic field theory.

All of this in turn was superseded by Einstein’s theory of special relativity (which made specific testable theories unseen or immeasurable via technology at that time, eg gravitational lensing – since proven).

Similarly Newton’s classic laws of mechanics have been succeeded by quantum mechanics, but his original theory still “works” at the macro-level – it only breaks down at the atomic level – in computing for example where quantum effects can often dominate.

There is also an economy to falsifiability: rather than stating all swans are white, and having to sample every swan on the planet to verify, the legitimate observation of a single, black, red, yellow or any other coloured swan would invalidate the theory.

It may still be correct to state “most” swans are white: but not all.

Major theories such as the standard model of physics or the theory of evolution by descent have had to withstand almost constant attempts at falsification.

Some well-intended, others less so.

The biologist JBS Haldane was once asked what experiment could over-turn evolution: his somewhat disparaging answer was “fossil rabbits in the Precambrian era”. In fact any single modern-day animal fossil would do. None have been found.

Creationists, on the other hand, have tried to counter the substantial evidence of the deep age of the Earth, in contradiction of traditional teachings, by invoking ideas such as the Omphalos Hypothesis, that Earth was made fully-formed by a creator who added all the appearances of age.

This hypothesis fails because it is itself unfalsifiable – unless a Creator appears and admits to this act, (an event well-worth being part of to be sure, when the Ultimate Being suddenly turns up to explain his prank): otherwise, no experiment or statement, even in principle, can be made to contradict it. It sits there as accepted or unaccepted fact, and generates no further prediction or observational forecast.

In Popperian terms it is not “science”.

For Popper the test of real science is that the idea or hypothesis or theory is ultimately falsifiable.

If it passes the valid tests it grows in strength, if it fails them it is superseded by a new, better theory.

Linus Pauling summarized this by noting that theories could be right (unfalsified so far), wrong (but at least testable) or, the poorest, not even wrong (unable to be tested).

Falsification and Elegance

It is worth admitting that even in the science community the concept if far from settled, and it induces deep controversy of almost theological proportions.

In one camp are the traditional science advocates, exemplified in a recent paper (here) that calls for the retention of these principles of predictions, testability and falsifiability as the best way to continue to develop scientific discovery and enquiry.

In the other, as Ellis and Silk in the paper discuss (from 2014):

This year, debates in physics circles took a worrying turn. Faced with difficulties in applying fundamental theories to the observed Universe, some researchers called for a change in how theoretical physics is done. They began to argue — explicitly — that if a theory is sufficiently elegant and explanatory, it need not be tested experimentally, breaking with centuries of philosophical tradition of defining scientific knowledge as empirical. We disagree. As the philosopher of science Karl Popper argued: a theory must be falsifiable to be scientific.”

They then take this issue into a wider context:

“This battle for the heart and soul of physics is opening up at a time when scientific results — in topics from climate change to the theory of evolution — are being questioned by some politicians and religious fundamentalists.“

So. to the core question posed: with the increasing pressure to react to the risks of climate change and the rapid emergence of solar, wind and battery technologies, are we at a sudden energy tipping point: a fast energy transition?

And is it a falsifiable theory ?

1. The Manufacturing of Energy: Is The Fast Transition Falsifiable?

That an energy transition in underway (and has to be underway – see above) is not systemically argued, although it’s pace and magnitude are disputed and denied in several quarters.

However – the phrase “energy transition” needs to be more tightly defined.

As per the previous “transitions” it does not mean the total substitution of one energy source by another.

It more accurately signifies that point at which an incumbent energy system has peaked in absolute demand, and enters long-term secular decline as the incoming energy system first consumes all marginal growth in the primary energy system, and then starts to dominate energy supply.

As soon as this happens, a new dynamic takes place: the new entrant starts to expand more rapidly as its cost base continues to fall, taking further market share, and stimulating equity and capital investors in the incumbent sector to exit, shifting capital toward the new growth sector.

The key point to note is that this happens when the new energy system is still at a low share of overall demand – perhaps no more than 5% – as long as it is growing fast enough to absorb marginal energy demand.

Thus the transition point can be reasonably defined as the point at which the new energy achieves all marginal growth and the incumbent system enters a peak in absolute demand.

The issue now is how quickly will this transition be in response to the current demands of the planet’s environmental requirements – both global climate change and increasing more local concerns around air quality and health impact.

Two broad schools have developed: the orthodox dominant view is that modern energy transitions take a long time (human-scale, not geological), perhaps 50-100 years: empirical examples are the move from biomass to coal in the 1800s for heating and power, and coal to oil (then gas) for light and transport in the 1900s.

The key principles here are incumbency and energy demand growth: the incumbent system is still used substantially even as the new entrant develops, and because the new system allows total energy use to increase due to efficiency or cost benefits there is only energy diversification or layering, not substitution.

Hence the “coal century” of the 1800s still relied on 85% biomass energy; the oil century of the 1900s relied on 40% of energy delivered by coal.

The second school suggests that the next transition will be far quicker and very different to those that have gone before.

This argument rests on three issues: first, there have only been a few global energy transitions to rely on so there is a very limited statistical or theoretical basis to make any precise forecasts whether pro or anti the speed of transformation – after all in terms of major energy transitions, N = 3, is hardly a strong statistical basis.

Second, this is the very first transition from extraction-based carbon fuels to a totally different energy technology. Previous (major) transitions have been from carbon fuel to carbon fuel: biomass to coal to oil to gas (note – in overall primary energy terms hydro and nuclear have been of modest impact at < 10% for most their decades-long presence).

This time it is from carbon fuels to manufactured, scalable, conversion technologies using the planet’s (limitless and free) solar and wind energy as the primary energy driver.

The new energy is different in kind to the current energy source, not in degree.

It is far closer to digital technologies than the bulk off-line engineering structures of  today’s centralised thermal fuel systems.

In this way the new energy system can access learning curve improvements that traditional extractive energy systems cannot: solar and wind costs have decreased by over 80% in the past 10 years due to these effects, whilst oil, gas, coal and nuclear energy costs have essentially remained flat.

Both trends are set to continue – the fundamental nature of extraction costs which increase over time as larger deposits run out, and manufacturing costs which decrease over time via learning by doing.

The impact here is quick and wide-ranging – see below for power generation, but similar fast trends are occurring in EV battery efficiency and the transport sector for example.

Third and finally, due to the urgency of climate change (see above), this transition will also be driven far harder by social license issues than previous ones.

That is, previous transitions were motivated by offering an unenergized civilisation the ability to access heat, light and mobility at affordable cost.

Only now is it clear what the longer-term planetary costs, the externalities, of the energy system required have turned out to be.

The current system, for all it previous success, now comes with three unsustainable side-effects: not only global climate disruption, but local and regional air quality and pollution issues as the world urbanises, and deepening energy security concerns – over 80% of the countries in the world import oil and gas from just a handful of exporters, dominated by OPEC, Russia and the US.

The new technologies of wind and solar (plus battery storage) present a far more economic, equitable and effective global energy system.

This energy transition is therefore not another slow-motion high-carbon to lower-carbon baton pass.

It is a way to fundamentally break free from the carbon extraction and burning cycle that has driven global energy access for two centuries, but which now threatens the very civilisations it has built.

As this an increasingly well-perceived phenomenon, this energy transition will be driven far faster as the socio-political issues compound the technological and economic drivers – a technological, cultural and political tipping point.

This then is the fast transition hypothesis:

The energy switch, from fossil fuels to manufactured, non-carbon, wind/solar/battery conversion technologies will be far quicker than previous energy conversions due to compelling technology and economics, intensified by social license issues driven by climate change, air quality and energy security concerns.

Which means that fossil fuel demand will likely decline in a rapid, S-curve or non-linear way, and cause the major disruption of the incumbent oil and gas industry almost immediately, ie in the early 2020s.

2. Falsifying The Fast Transition Hypothesis

It may seem inappropriate to subject a hybrid socio-economic technical idea to the rigour of empirical scientific analysis.

Indeed Popper was even against evolution as a falsifiable theory –  but eventually he recanted;  clearly, precise tests such as rabbit bones in the Precambrian improved its prestige.

So if we can expose climate science to scientific scrutiny, we should be able expose a technical energy system that is driven by climate urgency to the same objective tests as well.

Any fast energy transition “theory” therefore must have  exact, testable predictions and be falsifiable.

We suggest such predictions fall in to three categories: technology uptake, economic and investor impact, and policy and social response.

The Fast Transition hypothesis suggests that the uptake of renewable technology will be at the rapid end of current models, market reaction equally fast and social and policy response increasingly supportive of change.

A good detailed summary of this “disruptive” model of transition, and it’s alternate, business as usual (BAU) is outlined in this excellent piece here.

The following section takes a number of examples from this report.

2.1 Fast Transition (Disruption)

Technology Uptake

A sample of key technology predictions under this model is shown below:

  • The cost reductions of wind, solar and EV batteries will continue at more or less their current pace until 2025-30 (ca 10-20% pa); this means that by 2025 solar and wind costs will generally be $20-30/MWh, and battery costs $50-100/kWh
  • Wind and solar are now generally lower in cost for electricity (LCOE) than new fossil fuels, and will be lower in cost than operating fossil fuels by 2025.
  • Growth rate of wind, solar and EVs: by 2025 wind/solar capacity growth pa will be over 200GW, and EVs will command 30% total sales (6-10% total fleet).
  • Emerging markets will not follow the OECD growth patterns: they will “leap-frog” technologies; thus China especially and India will embrace wind, solar and EV growth – China is therefore expected to dominate wind, solar and EV market share due to its inherent manufacturing expertise, and both China and India demand for oil products and crude will peak at rates far below OECD per capita equivalents in the early 2020s.
  • Wind / solar will make up 15-20% of total primary energy by 2025 (ca 100EJ)
  • A series of major incumbent energy peaks will occur in the early 2020s: global ICE sales will peak this year (or perhaps did in 2018), the gasoline demand peak will be in 2020, global oil demand will peak in 2022 in absolute terms and enter secular decline, global primary energy will peak in 2025 at about 650EJ overall

Economic Impact and Investor Response 

As financial markets will react far faster to transitions than incumbents, a fast transition hypothesis predicts financial sector activity will have already mobilised, some 2-5 years ahead of generally obvious sector changes.

Markets invest in growth, and so anticipate decline (their own falsification approach – if a market is not growing can any expansionary argument falsify this approach?  –  if not, continue to divest).

Thus, over the past 2-5 years fossil fuel financial markets have produced the following outcomes:

  • Main coal sector shares in bankruptcy in 2013
  • Main fossil fuel European power utility sectors down range of 30-80%: other fossil fuel utilities such as US PG&E and Australia’s Synergy showing major losses / potential bankruptcies
  • Oil&Gas share of S&P 500 dropped from 14% to 4%
  • Oil&Gas major stocks: down 25% from 2017, E&P stocks down 45%
  • Oil&Gas service sector stocks: down 65% since 2017, with exploration supply chain stocks down > 80%
  • Major investor groups weighing the need to divest further and reduce remaining exposure to the sector
  • Main Auto sector stocks down 20% from 2017: split down 40% legacy OEMs, electric platform 10%

Expectations for the next 2-5 years:

  • Any new oil and gas investment incompatible with Paris 1.5deg C – increasingly known meme which will weigh heavily on sector. Pressure to stop any new long-term projects
  • 2020 onwards – demand for sharp increase in dividends / harvest vs growth mode
  • Investor retreat from oil and gas association – investment, sponsorship, support
  • 2025 – Oil sector share prices down 50% from today’s level
  • Oil sector EP and supply chain bankruptcies ongoing from 2020
  • Auto sector bifurcates into 2 clear groups: electric-dominated platforms and ICE-traditionalists: former with share price resilience, latter share price collapse by 2025 (> 50% drop)

Policy and Social Response

  • Increasing international response eg the Inevitable Policy Response campaign – new calls for change / CO2 impact assessment (mode 3) across total consumer chain
  • Social license issues increase in intensity

o   Oil and gas companies pressured to stop long-term investments (extinction rebellion / other NGOs)

  • Consumer demand for greater action – oil companies spilt into camps – harvest / progressive – versus resistant incumbents

o   Progressive some share price resilience, but downward trend (10-20% drop 2025)

o   Resistant incumbents see share price collapse (> 50% by 2025)

2.2 Potential Falsification – The BAU Model

The falsification of the predictions above rest on a broad series of business-as-usual (BAU) proposals: slowing technology advance, restrained investor action and diminishing social response.

They are exemplified in the IEA NPS scenario and most IOC base forecasts or scenarios.

For a slow or delayed transition, the following outcomes have to emerge, in falsification of the fast transition hypothesis.

Slowing Technology Advances

  • Wind / solar learning curves slow or stop and do not improve upon incumbent cost levels – somehow manufacturing learning halts, perhaps due to tariffs.
  • Wind/solar capacity growth slows or flat lines to 2030 and beyond: perhaps due to policy or political objections to land use.
  • There are no new fossil fuel or ICE peaks until the 2030s or beyond: oil and gas continue with year on year growth for at least the new 10-15 years; ICE cars resume absolute growth even with EV competition
  • Return-to-coal and return-to-nuclear: both technologies stage a renaissance either through policy support in specific countries (France, China, US) or breakthroughs in costs / pollution control. This could see both increase as a share of total energy demand in the 2020s, not a decline.

Restrained Investor Action

Despite investors currently removing exposure to the fossil fuel sector, it is feasible that sentiment could return:

  • Select investors focus on the sector to trade the cycle – expecting a bounce back increasing market values – thus oil and gas sector shares increase in value through 2025
  • Investor sentiment in ICE car dominant firms act as a hedge to the transition not happening as fast – non-EV investing car firms show relative stock price improvement through 2025
  • Stock sector improvements in utility sectors and fossil fuel supply chain sectors would also falsify the hypothesis

Diminishing Social Response

Perhaps XR and other activists are a passing fad, which will dissipate due to government policy, less aggressive weather, activist aging or incumbent lobbying.

This element of the fast transition seems potentially the hardest to predict and hence falsify as it is the newest phenomenon, and difficult to define in detail.

That said, at least three quantitative tests have been suggested:

  • Improved share price performance of major fossil fuel investment firms (Exxon . Chevron)
  • Less pressure on long-term investors in oil and gas – via social media / policy
  • Acceptance of continued long-term investment in fossil fuel technologies eg regular sanction of mega capital oil and gas projects outside OPEC and returning interest from the investment community

3. Conclusion

The Fast Energy Transition theory is predictive and clear enough to be falsifiable. Many tests and observations would undermine the theory if they were to be made. The theory, whilst still to be fully verified, is robust enough to handle falsification challenges.

In fact it seems easier to falsify the belief that the energy world is not changing, than it is to falsify the claim that it is.

Falsification is an economical way to test a proposition or theory, in this case a rapid energy transition, versus alternatives such as forecasts and narratives or scenarios.

Finding evidence that clearly contradicts the theory pinpoints the flaws or inconsistencies of the argument.

So far the fast transition theory holds up: it is the contra notion that the change will be smooth, linear and of long duration that has more and more falsifying evidence to contend with.

Each week brings more evidence of technological energy improvements in wind, solar and battery EVs and storage; along with the closure “earlier than expected” by the fossil fuel industry of coal and gas plants.

“Earlier than expected” and similar phrases are the ongoing falsification of the  BAU theories happening in real-time.

Epilogue: Falsifiability and Scenarios

Some oil companies or analysts maintain they cover both disruptive and conservative models via various scenarios that offer quite different visions of the future.

However these scenarios do not offer the same intensity of challenge that a falsification hypothesis would contain.

Typically scenarios blend several characteristics of the two models, and tend to avoid extreme event-driven, tipping points.

Their predictions are also often qualitative rather than quantitative.

They generally allow for different scenarios to run in parallel with others rather than create a binary stop-go dynamic.

Falsification offers a break with current theories, not a different forecast using the same assumptions.

For that reason falsification may provide a more quantitative and predictive approachand be more effective as a route to reinforcing or debunking change – especially disruptive change.

For example, when the IEA and other oil analysts or companies produce yet another correction to their prior forecasts eg for solar growth (below), or over-confident oil demand resilience, it is not merely a poor set of models being corrected, or just one scenario that can be replaced by another.

Source AukeHoekstra

It is an underlying theory about the energy world (BAU) that has been falsified: continuous ad-hoc defense of the theory will not do.

Linus Pauling would label BAU wrong, and suggest it ought to be rejected.

In its place the theory of fast transition provides a more robust series of empirically-backed statements that better describes the energy system transformation underway.

It may yet be falsified (as summarised above) – but so far it has withstood these tests.

That it is falsifiable at all points to the objective reality of the fast  transition underway at a time when it is most needed.

————-   ————-