In March 2023 the ifo Institute publishes an issue of its Schnelldienst (its rapid policy bulletin) under the title: "Deindustrialisation: bogeyman or necessary step in the structural change of the German economy?" The answer of most of the authors is relaxed. Bialek, Schaffranka and Schnitzer write: "No reason to panic." Müller's heading is simply: "Allow structural change." The energy crisis, they say, is no game changer. The change will be more a continuation of the structural change that was coming anyway than the trigger of an abrupt deindustrialisation.
In December 2025 the ifo winter forecast reads: "Structural change has Germany firmly in its grip." Production potential is revised downwards by 0.7 percentage points. "Without structural reforms, further erosion of Germany as a business location looms."
In April 2026 ifo President Clemens Fuest writes in person: "The deindustrialisation now emerging in Germany is not only dangerous for growth as a whole; rural regions are also threatened with falling behind" — with "far-reaching economic and political consequences, up to and including the strengthening of populist parties on both edges of the political spectrum."
Three years. Three registers. The same reality — which has developed faster than forecast.
Energy crisis as an accelerator, not a game changer. Accompany structural change, do not prevent it. Skills shortage dampens the risk of layoffs.
Production potential revised sharply downwards. Industry keeps losing competitiveness. Special fund takes effect only with a delay.
Deindustrialisation dangerous for growth and regional cohesion. Danger of populism named explicitly. Reforms urgent.
I. What was said in 2023 — and why it sounded plausible
The reassurance thesis of 2023 was not stupid. It had three arguments, each containing a kernel of truth.
First: over the past decades Germany had coped remarkably well with the structural shift away from industry. The share of manufacturing in value added had fallen steadily from over 22 per cent to around 20 per cent — far less than in France (10 per cent) or Britain. This stability was a genuine argument for resilience.
Second: in autumn 2022 industry had cut its gas consumption by up to 27 per cent — by substituting other fuels, through efficiency gains, through short-term adjustment. That was impressive and spoke for adaptability.
Third: the skills shortage really did dampen the risk of layoffs. Anyone wanting to shed workers finds it easier to place them in a tight labour market. The demographic argument was real.
What the reassurance thesis underestimated — or rather did not underestimate but did not want to say — was something else. It treated energy prices as a cyclical problem in a structurally sound system. But the system itself was the problem.
II. What the economists did not say
The 2023 ifo Schnelldienst analyses the symptoms precisely. What it does not analyse: the structural cause of why Germany was so exposed.
Decades of cheap Russian gas were no gift of the market. They were the result of a political decision — deliberate dependence as an instrument of foreign policy. Nord Stream 1 and 2 were built against the resistance of the USA and Eastern Europe. The costs of this decision were always political, not technical. They fell due in 2022.
At the same time Germany had lost its ability to build on a large scale. Not to plan, not to finance — to build. The Zhanjiang paradox: in 2025 BASF opened an integrated Verbund site in China for €8.7 billion — completed without any notable delay. In Germany the expansion of the very same concept founders on permitting procedures, infrastructure deficits and energy costs. The comparison does not show Chinese superiority — it shows German paralysis.
And finally: the quiet erosion of the Mittelstand, Germany's backbone of small and medium-sized, often family-owned firms. 231,000 businesses without a successor by 2026. Not spectacular insolvencies — silent closures. A machine builder in Thuringia that was world market leader in a niche for 40 years closes when the founder retires and no buyer comes. That shows up in no crisis statistic. It is deindustrialisation all the same.
The 2023 ifo Schnelldienst did not ask these structural questions. It asked: is an abrupt deindustrialisation looming? Answer: no, rather gradual change. The follow-up question — whether gradual change in the wrong place is worse than abrupt change — was not asked.
III. What happened between 2023 and 2026
Energy prices did not return to pre-crisis levels. They remained structurally higher than in the USA and China — because the political decisions that led to them are not reversible. No more Russian gas. LNG more expensive. Renewables expanded, but not fast enough for industrial base-load demand.
The automotive industry — backbone of German industry, 17 per cent of industrial employment directly or indirectly — got caught between the pressure to electrify and Chinese competition. VW announced plant closures that were still considered unthinkable in 2022. Suppliers collapsed. The transformation to electric mobility, which had been sold as an opportunity, turned out to be an opportunity for Chinese manufacturers — not for German ones.
And insolvencies rose. Record figures in 2024 and 2025. 520,000 fewer industrial jobs than before the Covid period. The "gradual change" that economists had described in 2023 as manageable accelerated — because structures that must first break before they become visible break all at once when the next wave comes.
IV. The Megamachine reading: why economists warn too late
The ifo Institute's shift in tone between 2023 and 2026 is not a failure of individuals. It is structural — and therefore interesting.
Economic research institutes live in an ecosystem of commissioned research, political access and public influence. Reassurance is functional. Whoever says "no reason to panic" is invited to government briefings, quoted by ministers, booked for panels. Whoever says "the system is structurally sick and needs interventions nobody wants" loses mandates and is labelled "hostile to business".
That is not corruption. It is the logic of the Megamachine: institutions adapt their messages to what the system can absorb. In 2023 the political system could not absorb a radical diagnosis of deindustrialisation — because the answer to it would have required decisions no government wanted to take: accept structurally higher energy prices or subsidise them massively; pursue an active industrial policy instead of letting the market rule; not reform bureaucracy but dismantle it.
So the economists wrote: "No reason to panic. Allow structural change. Accompany the transition." That was politically compatible. It was not wrong either — it was incomplete. The gap between what the data showed and what was said was the gap the system left open.
By 2026 the gap has grown too wide. Fuest himself writes about the danger of populism and rural erosion. That is the language of someone who can no longer reassure — because reality is too obvious. That is not progress in understanding. It is reputation management under the pressure of facts.
V. What is missing now
Fuest writes in April 2026: cutting bureaucracy, cheaper energy supply, tax relief. That is the usual reform list. It is not wrong — but it is the answer to the question "how do we improve the framework conditions" and not to the question "how do we rebuild the industrial capacity that has been lost."
This second question is harder. Because lost industrial sites do not come back when taxes are cut. Because skilled workers who have left do not come back when bureaucracy gets leaner. Because supply chains that have shifted to China, Eastern Europe and Mexico have an inertia of their own that tax incentives influence only marginally.
Germany can no longer build — on the scale and at the speed that new industrial settlement requires. That is the real diagnosis. Not too expensive, not too bureaucratic alone — no longer capable. Institutional capacity has eroded: planning authorities that are understaffed, engineering knowledge that has drained away, craft businesses that find no successors, infrastructure that has not been renewed for decades.
Between 2023 and 2026 the ifo Institute sharpened its diagnosis — but did not change its method. It continues to describe symptoms and recommend improvements to framework conditions. The Megamachine question remains unasked: why did the structures produce what they produced — and who profits from their not changing?
Who profits from German energy prices that are higher than American ones? American LNG exporters. Who profits from German industry moving to China? Chinese locations — and the German corporations that produce more cheaply there. Who profits from 231,000 Mittelstand businesses without a successor? Private equity funds that take them over at bargain prices.
No ifo Schnelldienst asks these questions. Not because the economists do not know them. But because the Megamachine does not ask these questions — it only answers the questions put to it. And the only questions ever put to it are those that do not call the system itself into question.