HVO100 and the Combustion-Engine Ban: When a Good Study Becomes a Bad Strategy

Why a clean fuel is not a business model — and how Rumelt’s test tells them apart

There is a particular move in public debate that I have come to find more interesting than the topics it is applied to. Someone produces a careful, narrow, technically honest piece of analysis. Then a second group picks it up, strips away the qualifiers, and presents it as proof of a sweeping claim the analysis never made. The original author rarely objects — partly because the louder version is flattering, partly because the gap between what was shown and what is now asserted is invisible to anyone who did not read the footnotes.

The current German fight over the combustion-engine ban is a textbook specimen. And because I spend my working life asking whether something is a good strategy or merely a good story, I want to use it as a worked example. Not to argue for or against synthetic diesel — I genuinely do not care, for the purposes of this piece, whether HVO100 is wonderful or worthless. I want to ask a narrower and, I think, more useful question: when does an analysis support a strategy, and when is it being used to disguise the absence of one?

For that question I lean on Richard Rumelt’s Good Strategy / Bad Strategy, which remains the sharpest instrument I know for telling the two apart. Rumelt’s central, uncomfortable point is that bad strategy is not the absence of good strategy. It is its own thing, actively produced, with recognizable fingerprints. The HVO100 debate has all of them.

What the HVO100 study actually measured

In May 2026, Professor Thomas Heinze of the Institute Automotive Powertrain at htw saar published an “Übersichtsstudie HVO100” — an overview study of hydrotreated vegetable oil, the paraffinic diesel substitute now sold at German pumps under the XTL label. As a piece of fuel chemistry and engine engineering, it is competent and, within its scope, fair. It documents that the HVO100 molecule burns cleaner than fossil diesel: a cetane number above 70, essentially no aromatics, lower particulate raw emissions, a friendlier water-hazard class, better storage stability. It reports a greenhouse-gas saving of up to 96 percent depending on feedstock. It notes that the fuel drops into existing diesel engines with no modification.

None of that is the problem. Every one of those claims lives comfortably inside the study’s chosen frame: the properties of a litre of fuel and what happens when you burn it in one vehicle.

The problem is what happens next — when the study leaves the laboratory and enters the argument.

The Fuel That Proves Nothing

Read how the study is invoked, and you will see the same structural jump every time. A trade body, the eFuels Forum, calls it “an important basis for a fact-based and technology-open debate” and positions it directly against the political focus on electric mobility. The “HVO100 goes Germany” campaign — run by an automobile club together with carmakers, fuel retailers and petroleum traders — turns it into the claim that HVO is “a gamechanger” and “a possible future for the combustion engine through technology-openness.” AUTO BILD reports that a new study shows the combustion engine is “considerably less superfluous” and that cars “would therefore not have to disappear.” Petitions against the ban cite “modern combustion engines running on HVO100” as the practical alternative. (Sources — eFuels Forum commentary: https://www.efuels-forum.de/2026/05/13/kommentar-zur-uebersichtsstudie-hvo100/ and https://www.efuels-forum.de/2026/06/09/neue-studie-hvo-warum-nicht-die-bestehende-dieselinfrastruktur-weiter-nutzen/ · “HVO100 goes Germany” campaign: https://hvo100.team/ ·

Notice the move. The study measured a molecule. The campaign asserts a system: that HVO can decarbonize the fleet, and therefore that the ban on combustion engines is unnecessary. The first is a fact about chemistry. The second is a claim about industrial scale, feedstock supply, opportunity cost and the entire value chain — none of which the study examined.

This is the precise point where a sound analysis is conscripted into a bad strategy. And Rumelt gives us the vocabulary to say exactly how.

Fingerprint one: the real bottleneck is feedstock — and it sits outside the frame

Rumelt insists that a real diagnosis names the one decisive constraint in a system — the chokepoint that limits everything downstream. A good strategy then concentrates force on that pivot point. The most common way analyses go wrong is not by getting the chosen link wrong, but by analyzing a conveniently reachable link while the actual constraint sits outside the frame entirely.

So where is the constraint for HVO?

Not at the pump, and not at the tailpipe. It is upstream, at the feedstock. And here the study is its own best witness against the campaign. Heinze reports, from the German agricultural agency’s 2023 figures, that HVO in Germany is produced almost entirely — 99.8 percent — from waste and residue streams: used cooking oil, category-2 animal fats, palm-oil-mill effluent. These are not abundant inputs that scale with demand. Used cooking oil exists only in the quantity in which the world already fries food; slaughter fats exist only in the quantity in which animals are already slaughtered. The supply is fixed by something other than how much diesel we would like to replace. (Source — figures from the German Federal Office for Agriculture and Food (BLE), Evaluations- und Erfahrungsbericht 2023, as cited in the htw saar study)

The numbers make the gap vivid. The total HVO energy base Heinze cites for Germany in 2023 is 615 terajoules. German diesel consumption is in the order of a thousand petajoules a year — more than a million terajoules. The waste-and-residue base from which “real” HVO is made therefore covers a fraction of German diesel demand that is far below one percent. You do not decarbonize a fleet of millions of vehicles with a feedstock that, today, covers a fraction of a percent of their fuel. (Source — Umweltbundesamt: German diesel share and transport energy consumption: Energieverbrauch und Kraftstoffe)

Heinze, to his credit, names the relevant tests. In his feedstock chapter he lists exactly the right questions a producer must keep asking: the quantity available, the prior use in other applications (“does residue use force other applications onto less climate-friendly inputs?”), and “last but not least the avoidance of land-use competition.” He even shows the RED III caps — used cooking oil capped at 1.9 percent of an obligated party’s energy volume, rising to 2.8 percent by 2039 — which are nothing other than the regulatory acknowledgment of scarcity.

But naming a test is not running it. He lists the constraint and then does not bring it home to the system question: is there enough feedstock to do what the campaign says HVO will do? The diagnosis stops one step short of the chokepoint. And that one withheld step is precisely the space the advocates move into. They read “cleaner molecule, drops into existing engines” and never reach the page where the supply runs out.

This is what Rumelt calls failure to face the challenge — and it is no less a failure for being committed by omission rather than by error.

Fingerprint two: when “could” becomes “is” — the conditional that gets laundered

There is a second, subtler fingerprint, and it is the one I find most instructive for readers far outside the fuel debate.

Watch the grammar. Wherever the study touches the scaling problem, it shifts into the conditional. The sustainable production base “can perspectively be expanded” through alternative feedstocks such as algae oil and jatropha oil. Third-generation fuels “must” play a role in future, “although their use is currently still small.” Production “can in future be increased” through algae and jatropha and “further as-yet-unexamined feedstocks.”

Every one of those is a could, not a did. Algae oil and jatropha have been the standing placeholders of the biofuel debate for fifteen years — jatropha had an outright investment bubble in the late 2000s that ended in land conflict and disappointment, and algae diesel has never become cost-competitive at scale. The conditional mood is doing enormous work here: it gestures at a solution to the supply constraint without committing a single fact.

And this is where the laundering happens. The author writes “could in future.” The campaign reads “is.” The conditional walks in through the front door of a university study and walks out the back as a present-tense promise. The reader who is paying attention sees a hypothesis; the reader who wants to believe sees a guarantee.

Here is the discipline I would urge on anyone — and it generalizes far beyond fuel. You cannot refute a conditional, because it asserts nothing. But you can operationalize it. When a strategy rests on “X could supply the gap,” do not argue back with “no it couldn’t.” Ask the questions that force the conditional to show its hand:

• Which projects exist today — in operation, not in a press release?
• What is the funded pipeline: site, capacity, commissioning date?
• Which of them have left the pilot phase and reached commercial scale? General feasibility has been known for years; that is not the open question. The open question is industrial volume.
• What tonnage do they deliver, measured against the gap that needs filling?
• And concretely: how do we intend to proceed from here — whose resources are committed, on what timeline?

Notice that this is simply Rumelt’s “committed resource allocation” criterion turned into a set of questions. A litmus test, not a counter-argument. A strategy that can name its pipeline survives the questions. A strategy that can only say “could” has, by its silence, already answered them. No accusation is required; you only have to ask.

Fingerprint three: a diagnosis reasoned backwards from the installed base

The deepest fingerprint is the one that explains the other two. Why does the constraint land so reliably outside the frame? Why does the conditional get such a warm reception?

Because the whole argument is reasoned backwards from the installed base.

Look at who carries the claim: petroleum traders, fuel retailers, combustion-engine carmakers, the automobile clubs that organize their customers. The slogan — “the engine is not the problem, the fuel is” — and the constant invocation of “technology-openness” are the rhetorical form of a preservation reflex. The diagnosis is not being drawn independently of what the actors already own. It is being shaped so that what they already own — the engine, the pump, the buried tank, the refinery, the combustion-engineering competence — gets to keep running.

Rumelt is blunt about this: existing routines and depreciated assets become the lens through which a problem is even perceived. The organization looks toward where its written-off assets stand, because it can use them “for free,” and so it overlooks both the real constraint and the possibility that the real lever requires building something new. Sunk cost and self-imposed scope are two faces of the same inertia. Good strategy is forward-looking — its value lies in what it produces in the future, not in what has already been invested — and it almost always means actively overcoming this path dependence.

A genuinely forward-looking diagnosis would ask: given a scarce, fixed supply of waste-derived feedstock, what is its highest-value use across the whole economy? Aviation and shipping, which have no battery option, are the obvious rival claimants. Burning a scarce residue in a passenger car that could run on a battery may be precisely the wrong allocation. But that question cannot be asked from inside a frame built to keep the existing fleet on the road. The frame forbids the question.

None of this requires anyone to be acting in bad faith. Rumelt is emphatic, and so am I: bad strategy is rarely the product of malice or stupidity. It is the product of inertia and of the unwillingness to choose and thereby disappoint someone. To choose is to forgo, and forgoing creates resistance — so bad strategy avoids the resistance by including everything and prioritizing nothing. “Technology-openness,” in this light, is not a strategy. It is the sophisticated-sounding refusal to have one.

Is the combustion-engine-ban fight uniquely German?

It is tempting, from outside, to file this as a peculiarly German episode, and the temptation is half right.

The debate is genuinely European. Germany, Italy and several other member states pressed the Commission to unwind the 2035 ban, and in December 2025 the Commission did soften it — replacing the outright phase-out with a 90 percent reduction target, the remainder to be offset with e-fuels, biofuels or low-carbon steel. So the argument is not Germany’s alone. (Source — European Commission softening of the 2035 phase-out, December 2025)

But its engine is. The e-fuels carve-out exists because Germany, at the eleventh hour in early 2023, withdrew its support for the zero-emission target unless a synthetic-fuels clause was added. And the reason is not a superior strategic diagnosis. It is the installed base, measured in jobs: by the ifo Institute’s count, more than 600,000 German jobs were tied directly or indirectly to the combustion engine. That is the path dependence, quantified. A country that has built a third of Europe’s cars, around the combustion engine, for a century, perceives the decarbonization problem through the asset it already owns. Which is exactly Rumelt’s mechanism, operating not on a firm but on a national industrial identity. (Sources — Germany’s 2023 e-fuels demand and the ifo job figure: https://www.sciencedirect.com/science/article/pii/S0301421524000429 · EU policy history and VDA positioning: https://europe.influencemap.org/policy/EU-Light-Duty-Vehicles-CO2-targets-3481)

The HVO study did not create that reflex. It was simply a convenient, credible artifact that the reflex could reach for. That is the quiet lesson about how these things actually work: the demand for the laundered claim comes first, and the study that can be laundered is found second.

What a real strategy would look like — and why its absence is the answer

Everything so far has been diagnosis: how a sound study is conscripted into a bad strategy. But debunking is the easy half, and it is the half I trust least to leave anyone better off. So before I generalize, let me finish the case honestly — not by saying what is wrong with the campaign, but by asking what a good strategy in its place would actually require. The fairest way to do that is to take the advocates at their word.

Suppose they genuinely believe synthetic and waste-derived fuels are the answer. Then their own argument names their first task for them. The binding constraint is production. Not the molecule — that is solved. Not the engine — it already runs. The chokepoint, by their own diagnosis, is whether enough fuel can be made. And a strategy that believes its own thesis concentrates everything on that pivot: it builds eFuel and HVO capacity at industrial scale, commits the capital, names the plants, the tonnage, the commissioning dates. That is what conviction looks like once it becomes a strategy rather than a slogan — Rumelt’s committed resource allocation, turned from a test you apply into an obligation they must meet.

So where are the investments? If the people carrying the “technology-openness” banner truly believed the fuel could decarbonize the fleet, the refiners and traders among them would be pouring capital into production today, because the prize would be the entire diesel market. If that capital is not flowing, only two readings remain, and both are uncomfortable for them. Either they do not believe their own thesis — in which case “technology-openness” is exactly the costume for an absent strategy that the installed-base reflex produces. Or they have run the numbers, and the business case for making enough fuel to feed the passenger fleet simply is not there — in which case they themselves have arrived at the unwelcome conclusion they set out to avoid: that the combustion car cannot be saved this way. The silence of the pipeline is not neutral. It is their own answer to the question they did not want to ask.

And this is worth saying plainly, because none of it is a verdict against the fuel. eFuels and HVO100 will get built — that is close to guaranteed. But they will be built where they are the only option, not where they are the convenient one: in aviation and shipping, which have no battery to turn to. That market already exists by law. EU rules now mandate a rising share of sustainable aviation fuel, climbing from a few percent today toward seventy percent by mid-century, and almost all of today’s supply is made from precisely the waste-and-residue feedstock the HVO debate is fighting over — at roughly three times the price of fossil kerosene, a premium a quota-bound airline pays and a motorist never will. The scarce residue already has a buyer who needs it more, can pay more, and has nowhere else to go. The fuel has a real future. It just isn’t in the tank of a car that could have run on electrons.

Five lessons for anyone judging a strategy

So let me leave HVO at the pump, where it belongs, and pull out what travels — because most of you reading this will never argue about diesel, but all of you in management will, this quarter, be handed an analysis and asked to build a strategy on it.

A sound analysis and a sound strategy are different objects, and the gap between them is where most bad strategy hides. A study can be impeccable within its frame and still support a conclusion it never tested. Your job, when someone hands you “the data supports X,” is to check whether the data and X share a frame — or whether X quietly enlarged the frame on its way out.

Find the chokepoint before you accept the diagnosis — and look for it outside the analyst’s chosen boundary. The most flattering analyses are the ones that optimize a link the organization already controls. Ask what the binding constraint of the whole chain is, and whether it sits in a part of the system nobody in the room is responsible for. If the analysis is silent on the real constraint, it is not a diagnosis; it is a description of the comfortable part.

Treat the conditional as a question, not a claim. “Could,” “can in future,” “has the potential to” — these are not arguments, and you cannot defeat them by argument. Operationalize them instead: which projects, what pipeline, what has left pilot, what tonnage, by when, with whose committed resources? A real strategy answers. A wish goes quiet. You do not have to call anyone a liar; you only have to ask for the pipeline.

Watch for diagnoses reasoned backwards from the installed base. When the proposed solution conveniently preserves everything the proposer already owns, be suspicious — not of their honesty, but of their frame. Path dependence does not feel like bias from the inside; it feels like realism. “Let’s not waste what we have” is the most natural sentence in business, and one of the most dangerous, because sunk assets are supposed to be irrelevant to a forward-looking choice and almost never are.

And remember the demand comes first. The laundered claim is not usually manufactured by the analyst. It is pulled into existence by an audience that needed it. When a conclusion is suspiciously popular with exactly the people it benefits, the interesting question is not “is the study wrong?” but “what did this audience need to be true, and did they go looking for an artifact that would let them believe it?”

These five teach you to see through a bad strategy. The sixth teaches you to test a proposed one — and it is the more demanding skill, because seeing through is destructive and testing is generative. It is also where the HVO case stops being about diesel and starts being useful everywhere.

The sixth lesson: does the alternative have a business model that holds?

Take the proposer at their word, then go to their bottleneck. When someone offers you an alternative — this is the way forward, not that — do not argue about whether it is appealing. Grant the thesis, in full, and then ask what their own thesis makes binding. Every proposed strategy names its own chokepoint the moment you believe it: if the fuel is the answer, the constraint is making enough of it; if the platform is the answer, the constraint is the users who must show up. Find that pivot and look for the committed resources flowing toward it — the capital, the hires, the plant, the dated plan. What you are really testing is whether the alternative has a business model that holds: a customer who pays the premium, an architecture that can deliver, a revenue model that closes. The committed resources are the tell, because they flow only when someone has done that arithmetic and found a model that pays. If they are flowing, you are looking at a real strategy and can judge it on its merits. If they are not, the proposer has answered you without meaning to: either they do not believe their own thesis, or they have quietly discovered it does not pay — and in both cases the absent pipeline is the verdict. You did not have to refute anything. You only had to notice what conviction would have built, and that it wasn’t there.

This is the move that completes the others, and it is the one I care about most, because the five before it leave you standing on cleared ground with nothing built on it. Diagnosis is Understand — necessary, and not enough. A description of what is wrong has never yet produced what should be right. Strategy is generative before it is analytic: its work is not to catalogue the present but to bring a better configuration into being, and that takes an act of imagination the diagnosis cannot supply for you. You have to dream bigger than the asset you happen to hold — and then act on what you see, by building the production your own thesis demands or accepting the conclusion that its absence is already writing. Understand. Dream bigger. Act. The first keeps you from being fooled. The second and third are the only ones that change anything.

The test that ties all of these together is Rumelt’s, and it is brutally simple. A good strategy chooses, and choosing creates resistance. If the strategy you are being sold pleases everyone — keeps every asset, offends no incumbent, asks no one to give anything up — you are almost certainly not looking at a strategy. You are looking at a description of the present, wearing the costume of a plan.

HVO100 may well have a real and useful role somewhere in the energy transition. That is a separate question, and an empirical one. But “the molecule is clean, therefore we need not change the system” is not a strategy. It is a conditional in a lab coat — and the moment you ask it for its pipeline, it goes quiet.

---

Sources

The study under discussion

• Heinze, T. (2026). Übersichtsstudie HVO100, Institut Automotive Powertrain, htw saar, May 2026. https://www.htwsaar.de/fileadmin/user_upload/Übersichtsstudie_HVO100_Prof._Heinze__IAP_an_der_htw_saar.pdf

How the study is being invoked

• eFuels Forum, commentary on the htw saar study (13 May 2026). https://www.efuels-forum.de/2026/05/13/kommentar-zur-uebersichtsstudie-hvo100/
• eFuels Forum, “Why not keep using the existing diesel infrastructure?” (9 June 2026). https://www.efuels-forum.de/2026/06/09/neue-studie-hvo-warum-nicht-die-bestehende-dieselinfrastruktur-weiter-nutzen/
• “HVO100 goes Germany” campaign (Mobil in Deutschland e.V.). https://hvo100.team/
• AUTO BILD, “Optionen für den Verbrenner: Studie macht Hoffnung.” https://www.autobild.de/artikel/neue-optionen-fuer-den-verbrenner-28588871.html
• AUTO BILD, “Ökodiesel HVO & XTL: Was bringt der Biokraftstoff?” https://www.autobild.de/artikel/hvo-100-oekodiesel-klima-20987829.html

Feedstock and consumption figures

• Bundesanstalt für Landwirtschaft und Ernährung (BLE), Evaluations- und Erfahrungsbericht 2023 (feedstock mix, energy base). https://www.ble.de/SharedDocs/Downloads/DE/Klima-Energie/Nachhaltige-Biomasseherstellung/Evaluationsbericht_2023.pdf
• Umweltbundesamt, Energieverbrauch und Kraftstoffe (diesel share of transport energy). https://www.umweltbundesamt.de/daten/umweltzustand-trends/verkehr/energieverbrauch-kraftstoffe

The “could” feedstocks — algae and jatropha, two decades of unkept promise

The article’s claim that algae oil and jatropha have been the standing placeholders of the biofuel debate rests on two well-documented boom-and-bust cases. Jatropha was promoted in the 2000s as a miracle crop that would yield fuel on marginal land without competing with food; projected plantings of 12.8 million hectares by 2015 never came close, expected yields failed to materialise, and the crop collapsed by the early 2010s — today a single large plantation remains, in Ghana. Algae followed the same arc: hundreds of millions in venture funding, repeated promises of commercial-scale fuel by dates that came and went, and no path to price parity with fossil diesel. Both remain perennially “almost there,” which is precisely what makes them so useful as rhetorical placeholders and so unreliable as a strategy.

Jatropha is not a credible escape hatch. It remains technically usable and regulatorily recognised, but the core problem that caused the earlier collapse has not disappeared: reliable, sustainable, low-cost feedstock at scale. The newer evidence points to pilots and renewed interest, not to a mature supply base.

• Mongabay, “Jatropha: The biofuel that bombed seeks a path to redemption” (2023) — boom-and-bust history; failed yield projections; the sole surviving large plantation. https://news.mongabay.com/2023/04/jatropha-the-biofuel-that-bombed-seeks-a-path-to-redemption/
• Greentech Media, “Hard Lessons From the Great Algae Biofuel Bubble” (2017) — venture-funded optimism, missed commercial deadlines, no price parity with petroleum. https://www.greentechmedia.com/amp/article/lessons-from-the-great-algae-biofuel-bubble
• Kant, P. & Wu, S., “The Extraordinary Collapse of Jatropha as a Global Biofuel,” Environmental Science & Technology 45 (17), 7114–7115 (2011), DOI 10.1021/es201943v — peer-reviewed analysis of the global jatropha failure. https://pubs.acs.org/doi/10.1021/es201943v

The installed base, measured in jobs

The figure for combustion-engine employment comes from a 2017 ifo Institute study commissioned by the VDA, the German automotive industry association. Working from the 2015 production structure, it counts at least 457,000 industrial jobs directly tied to combustion-engine products — engines, exhaust-treatment systems and the like — rising to more than 600,000 once indirectly coupled supplier and complementary sectors are included. It is worth noting where the number comes from and where it leads: the same industry-commissioned study that quantifies the jobs at stake also concludes that climate policy should be “technology-neutral” and argues against a ban. The figure that defends the installed base and the call for technology-openness come from the same source — which is the whole point.

• ifo Institute (Falck, Ebnet, Koenen, Dieler, Wackerbauer), Auswirkungen eines Zulassungsverbots für Personenkraftwagen und leichte Nutzfahrzeuge mit Verbrennungsmotor, commissioned by the VDA, ifo Forschungsberichte 87 (2017). Project page: https://www.ifo.de/projekt/2017-01-01/auswirkungen-eines-zulassungsverbots-fuer-personenkraftwagen-und-leichte — Full report (PDF): https://www.ifo.de/DocDL/ifo_Forschungsberichte_87_2017_Falck_etal_Verbrennungsmotoren.pdf

The European and German policy context

• Euronews, “EU carmakers to comply with 90% emissions reduction by 2035 as full combustion engine ban scrapped” (16 December 2025). https://www.euronews.com/my-europe/2025/12/16/eu-carmakers-to-comply-with-90-emissions-reduction-by-2035-as-full-combustion-engine-ban-s
• “Defueling the impasse: EU political discourse on e-fuels,” Energy Policy (the 2023 German position; the ifo figure of ~613,000 combustion-engine-related jobs). https://www.sciencedirect.com/science/article/pii/S0301421524000429
• InfluenceMap, “EU Light Duty Vehicles CO2 Targets” (policy timeline and industry positioning). https://europe.influencemap.org/policy/EU-Light-Duty-Vehicles-CO2-targets-3481

A note on the diesel-volume comparison: the study’s HVO energy base for Germany (615 TJ, 2023) is stated exactly. The German diesel figure is given as an order of magnitude (above 1,000 PJ) rather than a precise share, derived from the Umweltbundesamt data on diesel’s portion of transport energy consumption. The “far below one percent” claim holds comfortably across any reasonable reading of those figures.