A new economic theory of decision under uncertainty · governed options · ex-ante operationalisation · from the firm to monetary policy
At its heart, this is the creation of an economic theory of decision: every choice is an option in time, and uncertainty is not priced but governed. The same construction that replaces Real Options is what finally renders three foundational frameworks executable — Bellman for the first time in sixty-seven years, Markowitz carried into the firm’s own strategy as it never was after 1952, and Modigliani–Miller ex ante and by induction. From the single firm to economic and monetary policy.
One continuous line of work since 2005, from the valuation of inventory as a financial asset to a unified decision brain, Volvas. Real options were never real — the intuition behind them becomes usable only once decision itself is governed. And this is not theory alone: the framework is published and peer-reviewed, protected as intellectual property, and deployed in regulated practice.
The heart of this work is a theory: that every operational and strategic decision is an option in time, and that uncertainty can be governed rather than merely priced. Real Options Theory borrowed the language of financial derivatives and applied it where its assumptions do not hold. The work below shows — mathematically and conceptually — why that transfer fails, replaces it with a governed model, and demonstrates that the replacement is precisely what unlocks three foundational frameworks that had remained theoretical in applied practice.
The diagnosis
The model rests on the Wiener process — stationarity, independence, continuity, the Markov property. In strategic ramp-up, seed and growth, none of these hold. And conceptually there is no underlying, no seller, no counterparty, no price: the “option” is a spreadsheet assumption, not a contract.
The replacement
Every strategic decision is a governed exposure to a tested demand curve. The spot becomes expected market value (P·Q); the strike, full production cost; the premium, the maximum rational exposure to reach margin. Flexibility is not priced — it is governed.
Why it matters
Supplying a tested state space, admissible decisions, a feasibility boundary and a computable value function does more than fix Real Options. It provides exactly the inputs that three foundational frameworks always left implicit — and so renders them operational, in advance of time, for the first time.
Signals → Forecast Intelligence → CRC → {BOS · BFI · MRI · MRM} → Multi-state → Bellman → optimal path
Bellman
1957 · operational after 67 yearsBackward induction is a valid result that stayed operationally out of reach in applied finance for want of a governed state space, admissible decisions and a computable value function. Supplying all three, this work demonstrates — for the first time in sixty-seven years — that Bellman’s principle can be executed on a real strategic decision.
“Unlocking Bellman After 67 Years” · ICFT 2026Markowitz
1952 · multi-state, for the firmMean–variance selection — never carried into the firm’s own strategic choices since 1952 — is extended beyond the financial portfolio to same-objective, multi-state comparison through the Brutman Flexibility Index, which weighs a path by the flexibility it leaves open, not by its return alone.
BFI · ICFT 2026Modigliani–Miller
1958 / 1963 · ex ante & by inductionValid only at equilibrium, MM is made executable state by state via the Maximum Revenue Model under a double constraint — capital structure and real financeability ratios. Projected before the fact and re-solved by backward induction, it lets the optimal balance sheet — and the strategy itself — be revised as conditions change.
MRM · MRM-Debt · ICFT 2026The frameworks become executable only because a working engine supplies their inputs in real time. One layer keeps the forecast honest; four capsules turn it into a governed decision.
Forecast Intelligence — the layer that governs the trajectory
Forecast Intelligence is what turns a static projection into a self-correcting trajectory. Each day’s data tests the demand curve that justified the decision; statistical drift detection — z-score, χ², directional inversion — flags a deviation the moment it appears, before it ever reaches the financial statements, and triggers a review of the alternatives still open. Its uses run across the whole architecture: it sizes nothing on its own, but it tells the capsules when a forecast has stopped being true — and so when to re-decide. This is the difference between reporting a loss after the fact and preventing it in advance of time.
Brutman Option Signal
Tests whether a decision is feasible against its demand profile — contracted or not — and yields the firm’s value before the first unit of revenue. Uncontracted demand is read as a renewable option on the future, not as a dead end.
Brutman Flexibility Index
The capsule that makes future flexibility itself a criterion of decision. It measures how many viable, structurally coherent paths remain open, and weighs each strategy by the optionality it preserves — so a slightly less profitable but more reversible path can rightly be preferred. Flexibility is not priced; it is governed.
Maximum Rational Investment
Caps investment not at expected return but at what remains rational under a double constraint — structural viability and the firm’s real treasury and financing capacity. It answers “how much can we still commit?”, not “how much might we gain?”.
Maximum Revenue Model & MRM-Debt
Sets the revenue and the maximum debt compatible with real financeability ratios, and projects the firm’s value and risk — to the Altman Z-score — recomputed at every debt maturity. This is Modigliani–Miller made executable, balance sheet in hand.
Read together — Forecast Intelligence testing the curve, BFI holding flexibility open, MRI bounding exposure, MRM carrying value, risk and debt — the engine does not predict the future. It governs the decision while there is still room to change it.
One intuition — every operational and strategic decision is an option in time — carries, without a break in method, from the workshop floor to the central bank, and from practice into theory. The four movements that follow are radii of a single model. Each is set out in full below, and each carries a real consequence: a lighter information system, capital structure decided ex ante, an artificial intelligence finally able to decide, and a route into monetary policy and economic theory.
The unifying brain
Volvas — the logical brain that unifies these decisions.
It holds in one place what every framework left implicit: a tested state space, the admissible decisions, and a computable value function — and arbitrates them in real time. The same engine governs each of the four movements below, and its natural continuation is ERP Downsizing: the governed logic folded back into the company’s own information system.
The option-based approach unifies what used to be managed apart. Inventory, supply chain, pricing, marketing, R&D and innovation cease to be separate disciplines and become the governed choices of a single system — each an option in time, each tested against the same demand curve, each bounded by the same rational-exposure limit. Whether to hold or release stock, to advance or defer a campaign, to fund or stage a development is read on one instrument, with switching costs priced in advance rather than discovered after the fact.
This is also where the method corrects everyday errors that accounting quietly endorses. Buying cheaper is not always a gain when lead times are long; a margin rate means little until it is set against time and risk; a line that is unprofitable in isolation may be profitable to the whole; advertising and R&D are charges in the accounts but investments in substance — options on future revenue — just as a workforce is, or can be, an asset that creates value rather than a cost to compress. The engine measures these where neither bookkeeping nor a law firm can: BFI keeps the reversible path open, MRI caps what may rationally be committed, and Forecast Intelligence re-orients the trajectory before the accounting close.
The practical consequence is ERP Downsizing. Once decisions are governed by one engine rather than negotiated across functional silos, the heavy enterprise-software layer that existed to reconcile those silos has far less to reconcile. The governed logic is folded back into the information system itself — a lighter, faster decision architecture that does not require an eighteen-to-thirty-month ERP deployment to coordinate what the engine already coordinates. The decision brain does not sit on top of the software stack; it replaces much of what the stack was for. This is the operational horizon of the whole architecture — and the next system to be built.
Modules BFI · Forecast Intelligence · MRI · SMCH
Here the replacement of Real Options does its decisive work — and it is finance and strategy, not finance alone. Supplying a tested state space, admissible decisions and a computable value function, the model renders the three classics operational before the fact. Each strategic choice is priced as a governed exposure: the spot becomes expected market value (P·Q), the strike full production cost, the premium the maximum rational exposure to reach margin. MRI sets the maximum rational investment under a double constraint — structural viability and the firm’s real financeability. MRM projects the optimal capital structure together with the firm’s value and risk, to the Altman Z-score, recomputed at every debt maturity.
The consequence is that finance and strategy are decided together. An investment that looks attractive on IRR or NPV but is too risky for the balance sheet that carries it — for its very mode of financing — is seen as such before the commitment, not after it. Re-solved by backward induction, the optimal structure and the strategy itself are revised as conditions change. This is exactly what Real Options left out: it reasoned on a project in the abstract, blind to the health and the financing of the firm that bears it. The BOS, by contrast, fixes the maximum admissible exposure and integrates the mode of financing — precisely the dimension the older theory could not see.
It also answers a question classical finance cannot even pose: what is a decision worth before it is taken, and what is the best alternative to it? BFI compares every path that shares one objective, with its hypotheses and its pre-costed switching cost, and the firm chooses not the highest projected return but the path that creates value with the least irreversibility. Profitability and viability are weighed in the same act.
Modules BOS · BFI · MRI · MRM · CRC
The models do not merely use artificial intelligence; they supply what it is missing. A learning system can predict, but without a governed state space, admissible decisions and a value function it cannot truly decide. Volvas provides exactly these — which is why the architecture is, in the language of the patent, logically necessary to operationalise Bellman: remove any one of the four required inputs and the claim to execute backward induction fails mathematically. The contribution to AI is therefore structural, not cosmetic.
And because the same architecture governs the symbolic and the non-monetary alongside the financial — through BAM and CRC — it advances the reasoning of intelligent systems, not only their forecasts. The capsules treat a strategic choice and a non-monetary structure with one governed logic: Volvas has been used to draft and to revise works as far from a balance sheet as a stage play. The claim is precise, and it generalises — a brain that can govern a decision can govern a structure, financial or not.
In practice the intelligence enters as middleware. It coordinates decision matrices across functions, replacing conflicting departmental targets with a single value-creation logic, and it reads Forecast Intelligence signals to speak before a curve breaks — a guardian of coherence rather than a passive predictor. It does not replace human judgement; it preserves it, by surfacing the fork while there is still time to choose.
Modules Volvas · BAM · CRC · SMCH
At its widest radius the work reaches policy. A Monetary Theory of Governed Signals treats refinancing itself as a reflexive, cognitive system rather than a fixed rule — the same governed-option logic, applied to the money that funds the economy rather than the firm. Because that logic operates identically at the level of a single decision and at the level of their aggregate, it opens questions the literature has long left theoretical. They are pursued here as a research programme, developed in a forthcoming paper complementary to the monetary theory.
A governed account of how individual decisions — each an option under exposure — aggregate into macroeconomic outcomes. The passage between the two scales is constructed, not assumed: the same arbitration that sizes a firm’s exposure composes, when summed, into a macro trajectory that can be read and governed.
By tracking the tested demand curve and the exposure it justifies, the model makes marginal utility an operationally measurable quantity rather than a theoretical abstraction — computed on real data, at the moment of choice, instead of postulated. A century-old object of theory becomes something a system can actually evaluate.
Already raised in the monetary-theory paper, the framework sets out the conditions under which the two traditions might be read as compatible rather than opposed — coordination by governed signal reconciling the demand-side and the information-side reading of the same economy. It is posed as a question the work makes it possible to explore, not a result it proclaims.
These are not victories declared but a programme the work opens — the natural consequence of a model that governs decision continuously, from the single firm to the system that finances all of them.
Modules Monetary Theory of Governed Signals · forthcoming
Full texts are protected works held by their publishers. References below link to the publisher of record or the SSRN abstract. Complete papers are available on request.
Full texts are not posted here, in keeping with the rights held by Wilmott Magazine, Springer, IEEE and other publishers. To request a copy for research purposes, write to contact@brutmanresearch.org. The complete SSRN author record is available via ORCID 0009-0000-0655-1389.
Invited lectures and conference presentations in financial technology, decision science and computational economics.
The models are not only theoretical. They are deployed operationally within two regulated entities, which apply Forecast Intelligence, the capsule architecture and the Maximum Revenue Model in advisory and investment-banking work.
Advisory and regulated investment-banking activity applying Forecast Intelligence, MRM and the governed capsules to performance, treasury, restructuring and financing.
edda-performance.com ↗The same architecture applied to investment structuring and the projection of optimal capital structure, ex ante, across financing operations.
edda-stockfinance.com ↗The work is not only argued. It exists as registered intellectual property and as running software, with independent academic validation — which is what allows a theory to be deployed in regulated practice rather than merely cited.
Patent
The governed engine is the subject of a USPTO provisional patent application — Volvas Brain: a Governed Multi-Capsule Decision Architecture for ex-ante investment valuation, capital-structure optimisation, multi-period trajectory governance and universal symbolic extension. Its claims set out the engine and the four inputs that render Bellman’s backward induction executable, presented as logically necessary to that result.
Copyright
The framework and its software are protected by a consolidated body of U.S. Copyright registrations — anchored in a dated priority chain of works. Copyright protects the source expression independently of the patent, and the continuity of that chain since 2005 is itself the strongest protection against replication.
Software
The architecture is implemented as running code: a corpus of Python modules — document factory, broker connector, company brain, financing-cycle engine, token engine, SPV valuation — realising Forecast Intelligence and the governing capsules (BOS, BFI, MRI, MRM, CRC) as an operational system, not a diagram.
Peer review
The core claims have been validated through peer review: presented at ICFT 2025 (Hong Kong) and published in Springer CCIS, Vol. 2868. The publication establishes prior art for the BOS three-formula architecture, the BFI as an integral of governed options, the MRI, and the AI-as-middleware and Self-Managed Company Hub theses.
Thierry H. Brutman is a researcher and practitioner working at the intersection of financial decision theory, option-based modelling and applied artificial intelligence. His work is distinguished by an unusual continuity: the central intuition — that every operational decision is an option in time — was first published in 2005 and has been developed, formalised and operationalised across two decades, culminating in a unified architecture that renders Bellman, Markowitz and Modigliani–Miller executable ex ante.
He directs the EDDA Stock Finance Research Laboratory, which has established formal intellectual priority for the framework through a continuous body of copyrighted and published work. His research is applied in practice within regulated advisory and investment-banking activity.
He has taught the magisterial course in financial management at the University of Paris II Panthéon-Assas and lectures internationally as an invited speaker and keynote.
Correspondence
Complete texts are available on request for research purposes. For academic correspondence, collaboration or speaking enquiries:
contact@brutmanresearch.org