THE CASCADE EQUILIBRIUM
German defense procurement is not broken. It functions exactly as designed: every authority reviews, approves, safeguards. Every single decision is correct. And at the end stands an outcome nobody wanted. Classical game theory has no model for this. Here is one.
I. The Labyrinth
A soldier needs a radio system. Between the moment the requirement is identified and the moment the equipment is delivered, an average of 10 to 15 years pass in Germany. For complex weapons systems, it is 15 to 25.
This is not an exaggeration and not an isolated case. It is the system. The Bundeswehr's Customer Product Management (CPM) — the rulebook governing procurement — encompasses dozens of phases, hundreds of decision points, thousands of jurisdictions. Analysis phase, risk reduction, solution proposal, realization, deployment. In every phase: reviews, approvals, queries, revisions, further reviews.
Eighty-five percent of the time on a federal road project passes before the first shovel breaks ground. For defense projects, the ratio is even worse. The Bundeswehr began procuring the CH-53K heavy transport helicopter in 2011. In 2026, the contract has still not been signed. Fifteen years — and the aircraft isn't flying yet.
The standard explanation is: bureaucracy. The correct explanation is game-theoretic.
II. The Structure
Players: A chain of institutions I₁, I₂, I₃, ... Iₙ that decide sequentially. Each institution has its own mandate, its own incentives, its own risks.
Move: Each institution receives the output of the previous one as input and makes a decision: approve, modify, refer back, or block.
Incentives: Each institution is punished if it lets an error pass — but not if it delays the overall outcome. No player is rewarded for the end product reaching the user quickly.
Information: No player has oversight of the entire chain. Each sees only their step and the immediately preceding one.
The outcome: Every player acts rationally. Every individual decision is correct. The aggregate outcome is irrational.
This is not a coordination game in the classical sense. In a coordination game, players decide simultaneously and fail to align — like two pedestrians trying to dodge each other on a sidewalk and choosing the same direction. In the Cascade Equilibrium, players decide sequentially. And that is precisely the problem.
Nor is it a sequential game in the classical sense. In chess or a Stackelberg game, the second player sees the first player's complete strategy. In the Cascade Equilibrium, each player sees only the intermediate result of their predecessor — not the chain of decisions that produced it, nor the consequences for the successor.
And it is not an information cascade in the sense of Banerjee and Bikhchandani, where actors abandon their private information in favor of observed behavior. Here there is no private information being ignored. Here everyone acts on the basis of their own, accurate information — and still produces a wrong outcome.
The Cascade Equilibrium describes something new: emergent irrationality from sequential rationality.
III. The Mechanics of Self-Protection
The key to the Cascade Equilibrium lies in the incentive structure of each individual link in the chain.
Consider the civil servant at the Federal Office of Bundeswehr Equipment, Information Technology and In-Service Support (BAAINBw). Their mandate: ensure technical requirements are met. Their career depends on not letting errors pass. Their risk from delay: zero. Their risk from an error: parliamentary investigation, media exposure, career destruction.
The rational strategy is clear: when in doubt, refer back. Request another review cycle. Obtain an additional expert opinion. This costs nothing — at least nothing that is attributed to the civil servant. The delay is borne by others: the soldier waiting for the equipment. The unit operating with outdated materiel. The taxpayer paying inflation costs.
Now multiply this rational self-protection strategy by the number of authorities in the chain. Requirement identification in the Army. Planning in the Planning Office. Technical review at BAAINBw. Budget approval in the Ministry of Defense. Procurement law compliance by the contracting authority. Parliamentary oversight by the Budget Committee. Audit by the Federal Audit Office. Appeals by unsuccessful bidders before procurement tribunals.
Every single link adds a safety loop. Every safety loop is individually sensible. The sum of safety loops is a system that produces nothing except safety — against everything except failure through slowness.
IV. The 25-Million Threshold
German procurement law contains a threshold that makes the Cascade Equilibrium particularly visible: for contracts exceeding 25 million euros, the Bundestag Budget Committee must approve. This parliamentary oversight is democratically correct. A parliament should decide on major expenditures.
But the threshold creates a perverse incentive. Project managers learn to structure contracts to stay below 25 million — not because this makes technical sense, but because it avoids the parliamentary loop. A project that should cost 80 million as a whole and should be procured as a whole is split into four tranches of 20 million each. Each tranche must be separately planned, tendered, and awarded. Total costs increase. Delivery time extends. But the project manager has minimized their risk.
The 25-million threshold is a textbook example of the Cascade Equilibrium: a rule that is individually sensible (parliamentary control) produces, in combination with the incentives of subsequent links (risk avoidance by project managers), an outcome nobody wanted (more expensive, slower, fragmented procurement).
V. The Gold-Plated Solution
German defense procurement exhibits a phenomenon insiders call the "gold-plated solution": the tendency to specify and optimize every system until it is technically perfect — and practically unaffordable.
Game-theoretically, the gold-plated solution is a direct consequence of the Cascade Equilibrium. Why? Because every link in the chain has an incentive to add requirements — and none has an incentive to reduce them.
The soldier wants the best equipment. The planner wants to cover all contingencies. The engineer wants the latest technology. The reviewer wants to ensure every requirement is met. The parliamentarian wants to guarantee taxpayer money is spent on the best possible product. Nobody in the chain says: "This is good enough." Because "good enough" is a risk — the risk that someone asks why the better option wasn't chosen.
The result: systems with 300-page specifications that require 15 years of development and end up costing triple the original budget. The Puma infantry fighting vehicle is the textbook example: conceived as the Marder's successor, specified as a Swiss Army knife, delivered decades late and billions over budget. During a 2022 NATO exercise, all 18 deployed Pumas broke down. Technically ambitious. Operationally a disaster.
Ukraine shows the contrast: systems that are "good enough" — drones costing a few hundred euros, modified civilian vehicles, improvised artillery solutions — accomplish more in the field than the gold-plated solutions still stuck in the specification phase in Koblenz.
VI. The Consultant Republic
After the Cold War ended, the Bundeswehr shrank from 500,000 to 180,000 soldiers. In the administration, not just positions but expertise vanished. The gap was filled with external consultants — McKinsey, Accenture, IBM, PwC. The Defense Ministry at times spent over one billion euros annually on consulting services.
Within the Cascade Equilibrium, consultants exacerbate the problem in two ways.
First: consultants have no incentive to simplify the system. Their business model is based on the system's complexity. The more complicated the procurement, the more consulting is needed. A consultant who simplifies the system to the point of making themselves unnecessary acts against their own interest. In the Cascade Equilibrium, the consultant is not part of the solution — they are an additional link in the chain that extends the cascade.
Second: consultants replace institutional knowledge with transactional knowledge. A civil servant who has managed procurement processes for 20 years knows the informal shortcuts that keep the system functional despite its complexity. A consultant who comes for 18 months knows the rules — but not the exceptions. And when they leave, they take their knowledge with them. The system loses memory.
The result: an apparatus that consults itself without understanding itself.
VII. The European Multiplication
What produces the Cascade Equilibrium at the national level multiplies exponentially at the European level.
Twenty-seven member states with 27 national procurement systems, 27 national defense industries, 27 national security interests. The European Parliament estimates annual efficiency losses from fragmented procurement at 18 to 57 billion euros. The EU Commission proposed centralization in 2025 with its Readiness 2030 initiative. The European Council rejected it — in favor of intergovernmental mechanisms that perpetuate precisely the fragmentation costing 57 billion euros annually.
The Cascade Equilibrium explains why. Every member state acts rationally: France protects its defense industry. Germany protects its procurement procedures. Italy protects its jobs. Sweden prefers NATO mechanisms over EU structures. Every individual position is rational for the respective state. The sum is a continent spending over 300 billion euros on defense and still unable to defend itself.
The European cascade has an additional element the national one lacks: the veto loop. In intergovernmental mechanisms, every state can block. This means the cascade has not only sequential extension (each link adds time) but also lateral blockade (each link can halt the entire chain). The combination of both — sequential and lateral — makes European defense cooperation perhaps the purest example of a Cascade Equilibrium in the present day.
VIII. Why Reform Fails
The Cascade Equilibrium has a property that makes it particularly resistant to reform: it has no culprit.
In the Prisoner's Dilemma, you can identify the defector. In the Chicken Game, the reckless driver. In the Dependency Game, the patron deepening the dependency. In the Phantom Player Problem, the generation living at the next one's expense.
In the Cascade Equilibrium, no single actor acts wrongly. The civil servant at BAAINBw acts correctly. The parliamentarian on the Budget Committee acts with democratic legitimacy. The auditor acts according to their mandate. The procurement lawyer acts in legal compliance. Nobody makes an error. The system still produces an outcome nobody wanted.
Because there is no culprit, there is no point of attack for reform. Every reform proposal is blocked by the link whose jurisdiction it touches. "Abolish parliamentary oversight?" — Impossible, undemocratic. "Reduce technical reviews?" — Irresponsible, quality risk. "Simplify procurement procedures?" — Illegal, corruption risk. "Eliminate consultants?" — Then the expertise is missing.
Every line of defense is individually valid. In aggregate, they are the reason the system cannot reform itself: because every link has a rationally grounded claim to function exactly as it functions.
IX. Ukraine as Counter-Model
What happens when the cascade collapses?
Since February 2022, Ukraine has operated a procurement system that contradicts the German one in almost every respect. Decisions are made in days rather than years. Specifications are written by users, not planning staffs. "Good enough" is not a risk but a doctrine. Civilians modify drones in garages. Software developers write artillery apps at hackathons. The feedback loop between the front line and procurement is so short that improvements are implemented in weeks rather than decades.
This is not a model for normal circumstances. Ukraine operates under existential pressure that forces shortcuts unacceptable in peacetime — corruption risks, quality deficiencies, insufficient documentation.
But Ukraine shows what happens game-theoretically when the cascade is reduced to a single question: does it work at the front? When the outcome is the standard rather than the process, the Cascade Equilibrium collapses. Because suddenly a criterion exists by which every link in the chain is measured — not by its own correctness, but by the aggregate result.
Germany does not have this criterion. In peacetime there is no front line. There are only processes. And processes optimize themselves — not the outcome.
X. The Model
1. Sequential rationality. Players decide one after another, not simultaneously. Each player acts rationally within their mandate.
2. Local visibility. No player has oversight of the entire chain. Each sees only their step and the immediate result of the predecessor.
3. Asymmetric punishment. Errors are punished, delays are not. This generates a systematic bias toward self-protection at the expense of speed.
4. Cumulative irrationality. The aggregate outcome is irrational even though every individual decision was rational. The irrationality emerges not in a single step but as an emergent property of the chain.
5. Reform resistance. Because no single actor acts wrongly, there is no point of attack for reform. Every link defends its function with legitimate arguments.
The Cascade Equilibrium complements the two preceding models. The Dependency Game describes a player who sees the trap and cannot escape. The Phantom Player describes a game where the loser is not at the table. The Cascade Equilibrium describes a game in which there is no loser at all — only a system that loses.
And that is precisely what makes it perhaps the most dangerous of the three games. In the Dependency Game, there is at least an actor who could theoretically choose the exit. In the Phantom Player Problem, there is at least a generation that inherits the problem and can react. In the Cascade Equilibrium, nobody owns the problem. It belongs to everyone — and therefore to no one.
The only force that breaks a Cascade Equilibrium is a crisis so acute that it simultaneously changes the incentive structure of every single link. A war. An economic collapse. A system failure so spectacular that no link in the chain can still say: "My step was correct."
Germany has been waiting for that. For 30 years.
Ten to fifteen years from requirement identification to delivery. Eighty-five percent of the time before the first shovel breaks ground. All 18 Puma infantry fighting vehicles broke down at the NATO exercise. Eighteen to 57 billion euros in annual efficiency losses from fragmented procurement. And not a single culprit — because everyone decided correctly. This is not bureaucratic failure. It is an equilibrium. It means: everyone optimizes their step. Nobody optimizes the outcome.