In April 2026 Der Spiegel reports on a storm gathering in the American financial system. Private credit funds — shadow banks that lend without a banking licence — have reached a volume of $1.8 trillion. More and more investors want their money back. The funds are blocking withdrawals. Rating agencies are downgrading. The president of Blackstone blames rumours. The head of JPMorgan warns of hidden losses. A Yale professor writes of a “bank run in slow motion”.
The same pattern that could be observed in March 2008, before Bear Stearns collapsed.
This text uses the case of private credit funds not as a financial analysis but as a case study for three theses of the Megamachine paper. The case is particularly revealing because it shows how the Megamachine grows not despite regulatory intervention but through it — how the attempt to close the loop merely relocates the loop.
I. Positive feedback relocates — it does not disappear
After the financial crisis of 2008, regulators considerably tightened the requirements for bank lending. That was an understandable reflex: uncontrolled credit expansion had brought the system close to collapse; so a stronger negative feedback was to be built in, a braking mechanism that would prevent banks from lending money too easily.
The result was not what was intended. The demand for credit did not disappear. It shifted. Companies that could no longer get bank loans turned to private credit funds. These funds are not subject to the capital requirements of banks, nor to liquidity rules, nor to disclosure obligations. They are less regulated because formally they are not banks — although they perform the same economic function.
That is the mechanism of positive feedback in the Megamachine: capital always seeks the path with the least negative feedback. When regulation blocks this path, a new path emerges — with even less negative feedback than before. Regulation did not break the loop. It relocated the loop into less regulated, less transparent, less observable territory.
The volume of private credit funds has grown from marginal to $1.8 trillion — growth that took place in exactly the period in which bank regulation was tightened. That is no coincidence. It is the direct consequence of a regulation that narrows the walls without closing the roof.
II. The missing observer — one controls what one can measure
A simple closed control loop works without an observer — it directly measures what it is supposed to control. The thermostat measures the room temperature and controls the heating without having to estimate the internal state of the house. The observer is a higher stage: it is needed when the internal state of a system cannot be measured directly but can only be inferred from its outputs. If the observer is missing in such a system, the system runs blind — it controls outputs without understanding what is happening inside.
The Wharton professor Itay Goldstein, one of the financial crisis experts quoted in Der Spiegel, says a sentence that describes everything: “The system has grown very quickly, and there are parts of it we do not yet understand well.”
That is the diagnosis of the missing observer. What grew was not measured, because no one knew how to measure it. The interconnections between private credit funds, private equity firms and insurers are not transparent. The risks are not reported on a consolidated basis. The IMF issued a warning in November 2025 — not because the data were clear, but because the missing data are themselves a warning signal. Lack of transparency is the absence of the observer.
That is a structural parallel to 2008. Back then, too, the argument of the reassuring managers was: we see no systemic risk. That was true — because the instruments for measuring systemic risk were missing. The CDO system was opaque. The interconnections between institutions were not visible. The observer was missing. And when the system tipped over, the surprise was real — not feigned. People really had not seen what was coming, because they had not built what would have been needed to see it.
The financial system is highly interconnected. If weaknesses exist in one part of the system, they can spread to other areas. — Richard Berner, NYU Stern School of Business, former chief economist of Morgan Stanley
A D-component in the control loop — a response to rates of change rather than to actual values — would have reacted to the speed at which private credit funds were growing, long before the funds had become large enough to be systemically dangerous. The D-component is missing. No one intervened at $200 billion. No one at $500 billion. No one at $1 trillion. Now it is $1.8 trillion and the first dominoes are falling.
III. Capital extraction through ownership architecture — the opposite of the GmgV
The most revealing aspect of the private credit complex is neither the volume nor the opacity. It is the ownership architecture.
Large private equity firms — Apollo Global Management, Blackstone, KKR — have systematically bought insurance companies in recent years. The reason is precise: every year insurers receive premiums from millions of people. These premiums have to be invested somewhere. If the private equity firm owns the insurer, it can decide that the premiums flow into its own private credit funds.
The result is an ownership structure in which the money of pensioners and life insurance policyholders becomes the source of funding for the returns of private equity owners. The people who paid their retirement savings into a life insurance policy are not the owners of these funds. They are their funding base. Their capital is bound — but not for their benefit.
That is the structural opposite of the company with bound assets (Gesellschaft mit gebundenem Vermögen, GmgV). The GmgV binds assets so that they cannot be extracted — the capital stays in the company, serves the company's purpose, cannot be distributed to owners. Private equity insurers bind the capital of third parties — the policyholders — and divert it into the return structures of the owners. The binding serves not the purpose of those who built up the capital but the purpose of those who extract it.
The IMF calls it “potential conflicts of interest and a lack of transparency”. That is the most precise institutional formulation available for what the Megamachine paper calls structural capital extraction: an ownership arrangement that inserts between capital and its nominal owner an extraction apparatus that is legally legitimised and scarcely captured by regulation.
IV. The Megamachine at work — a system log
The private credit complex is not the failure of individual actors. It is the foreseeable result of a selection environment that systematically favours certain kinds of behaviour.
Step one: regulation after 2008 tightens bank lending. This is the attempt at negative feedback on the accumulation of risk. Selection pressure arises against risky bank lending.
Step two: capital seeks the path of least resistance. Private credit funds emerge as a regulation-free zone. They offer higher returns because they take higher risks — risks they do not have to disclose fully. Selection pressure favours growth in the unregulated area.
Step three: private equity firms realise that insurers offer a stable funding base. They buy insurers and divert the premium flows into their own funds. Selection pressure favours the vertical integration of funding source and return vehicle.
Step four: the system grows in a regulatory blind spot. No observer measures the overall size, the interconnections, the systemic risks. Selection pressure works against transparency, because transparency enables regulation that endangers the business model.
Step five: the system is too large to be ignored, but the instruments for measuring and steering it are missing. Regulators begin to warn — but without access to the relevant data the warning remains abstract.
No one planned this sequence. No one decided that the pension capital of millions of people should become the source of funding for private equity returns. It is the result of a thousand individually rational decisions made within a selection environment that favours certain kinds of behaviour and punishes others. That is the Megamachine: no will, no plan, no centre — only selection conditions that produce a result no one wanted and everyone created together.
V. What this means for institutional architecture
The case illustrates a fundamental weakness of the regulatory approach: regulation that covers individual institutions without covering the overall architecture shifts the problem instead of solving it. Banks are regulated more strictly — shadow banks emerge. Shadow banks are regulated more strictly — the next lightly regulated zone emerges.
That is the institutional equivalent of the arms spiral: every regulation produces an evasive reaction that requires a new regulation. The system as a whole does not become more stable, it becomes more complex — and thus harder to observe, harder to understand, harder to correct.
The way out of this spiral lies not in stricter regulation of individual institutions but in changing the ownership architecture itself. The GmgV is not a regulatory instrument — it is an ownership model that structurally excludes the logic of extraction, not by prohibition but by construction. Bound assets cannot be diverted into private credit funds, because there are no owners who could divert them.
That is the difference between closed-loop control and open-loop control described in the cybernetics chapter of the main paper: regulation is open-loop control — it sets inputs without measuring and adjusting the output. A changed ownership architecture is closed-loop control — it changes the selection conditions themselves, so that the undesired output does not have to be prevented by intervention but is simply no longer produced, because of the structure itself.
The price of lacking this architecture is concrete, and it is not borne by those who caused the problem. If the private credit complex collapses, it is not primarily the private equity owners who are affected. They have already realised their returns. Those affected are the people who paid into a life insurance policy for their retirement — and who did not know, and could not have known, that their premiums flowed into a system they would never have understood even if it had been explained to them.
This greater interconnection means that a collapse would not remain confined to the financial markets. It could affect the labour market, economic growth and household savings. — Itay Goldstein, Wharton School
That is the Megamachine in its most complete form: a system that transfers risks from those who take them to those who did not, through ownership structures that are legal, complex and completely opaque to those affected. No guilt, no intent, no possibility of accountability — only architecture.
VII. Chimeras — capital that no longer finds value creation
Why do such constructions arise at all? The answer lies not in the greed of individual actors but in the system logic of over-accumulated capital.
Capital that is already large enough seeks returns. Returns in value-creating companies require innovation — uncertainty, patience, the risk of failure. That is structurally incompatible with the time horizon and risk tolerance of capital that is already too large to wait for value creation. So the alternative emerges: constructions that simulate returns without generating them. They shift capital instead of increasing it.
Private credit funds are such chimeras. From the outset they have no potential for value creation — they are capital transfer machines. The money of pensioners and life insurance policyholders travels through a chain of ownership structures into the pockets of financial jugglers. In the end wealth has shifted without anything new having been created. A company that dares to innovate, opens up a market, creates jobs, develops products that did not exist before — that is value creation. A fund that collects money, lends it on at higher interest and charges fees in the process — that is redistribution with the appearance of productivity.
That is Mumford's Megamachine in its purest financial form: a system that claims to be productive and is in fact merely extractive. The complexity of the construction — the nested ownership structures, the opaque interconnections, the legally impeccable contractual frameworks — does not serve efficiency. It serves to make the transfer invisible.
And in the end no one is to blame. That is not cynicism — it is the precise description of the system. The private equity manager buys the insurer: legal. He diverts the premiums into his own funds: legal. The fund lends to risky companies: legal. The pensioner loses part of his retirement savings: a result of the market. No one has stolen. No one has lied. No single actor wanted or planned the overall result. The architecture produced it — through a thousand locally rational and legally impeccable decisions which, taken together, produce a result for which no one is responsible.
That is the strongest argument for changing the ownership architecture — not as a moral demand but as a structural necessity. As long as capital can legally be extracted, it will be extracted. Not because people are evil. Because the selection conditions of the system favour those who extract and disadvantage those who invest. The GmgV is not a regulatory instrument that punishes extractors — it is a form of ownership that makes extraction structurally impossible. That is the difference between open-loop and closed-loop control: not controlling the output, but changing the conditions under which the output arises.
VIII. Collapse as the final extraction phase
There is one last dimension of the private credit complex that is analytically uncomfortable but belongs to a complete description of the system.
The fund managers have already realised their profits — in the form of management fees, performance bonuses, dividends from their own insurers. The investors' capital is invested and illiquid. In this constellation a structural incentive arises that economics calls moral hazard: if the construction collapses, the manager bears a small part of the damage — the investors the large part. If the state intervenes to prevent a systemic collapse, the manager profits from the rescue. The risk is distributed asymmetrically — at the expense of those who provided the capital.
In this logic a spectacular collapse can be rational. Not as planned sabotage — but as the result of decisions that are locally rational under the given incentive structures. Whoever knows that a collapse will be read as a natural disaster has less reason for caution. Whoever knows that the narrative of the irrational run is available — the investors lost their nerve; without this run everything would have turned out fine — has an instrument of retrospective exoneration.
Jonathan Gray, president of Blackstone, has already put this narrative in position: “rumours in the media” are responsible, he says, for investors wanting their money back. That is not necessarily a lie — it is the preparation of a story. It was not the construction that was wrong. The investors lost their nerve. Had they kept their confidence, everything would have worked out perfectly.
The parallel to Bear Stearns is most precise here. In March 2008 the chief executive of Bear Stearns stressed that the balance sheet was strong despite speculation. A few days later the firm collapsed. The management had not launched rumours — it had offered a story that framed the collapse as an external disruption rather than an internal structural weakness.
The Megamachine analysis permits a precise formulation here: the system architecture creates incentives that make this behaviour rational — regardless of whether anyone acts deliberately in an individual case. That is the decisive point. Whether a manager consciously provokes a run or not is secondary. What is decisive is that the selection conditions of the system favour those who can use the collapse for themselves — and punish those it hits.
The collapse is not the failure of the Megamachine. It is its final extraction phase. The profits were privatised. The losses are socialised — through damage to retirement savings, through state bailouts, through job losses in the real economy. No one is to blame. The architecture was built so that no one can be to blame. That is its actual design principle.