The principal and the agent

A financial model is not just a calculation tool. It is a trust instrument.

This is not a metaphor. It is the central argument of Theodore Porter's Trust in Numbers — a 1995 study of how quantification became the dominant language of institutional decision-making in modern Western societies. Porter's observation is precise: in low-context professional cultures, where personal relationships and institutional reputation are insufficient to establish trust between strangers, numbers do the work instead. The budget, the projection, the financial model — these are not just tools for organising information. They are substitutes for the personal trust that the relationship has not yet had time to build.

But numbers only function as trust substitutes under one condition: they must be auditable. A number that cannot be traced back to its assumptions, that cannot be questioned, that produces outputs nobody can verify — that number is not a trust substitute. It is a trust request. It is asking the investor to believe without giving them the means to evaluate.

The crystal ball — in either of its forms — is a trust request dressed as a financial model. And in institutional finance, trust requests without audit trails do not get funded. They get filed.

On Agency theory.

To understand why, it helps to look at the relationship through the lens of agency theory — one of the foundational frameworks of institutional economics.

The developer who presents a financial model to an investor is operating within a principal-agent relationship. The investor is the principal — the party with the capital and the decision-making authority. The developer is the agent — the party seeking to deploy that capital on the principal's behalf. The relationship is structural and both sides know it.

The defining feature of any principal-agent relationship is information asymmetry. The agent knows things the principal does not — about the project, the risks, the assumptions behind the numbers. The principal cannot observe perfectly what the agent knows or how they are using it. This asymmetry is unavoidable. What varies is what each party does with it.

A transparent financial model — inputs clearly marked, assumptions visible, formulas traceable, outputs earned rather than asserted — is a voluntary reduction of the information asymmetry. The developer is saying, through the structure of the model itself: I know you hold the power in this relationship, and I am giving you the tools to evaluate me fairly. You do not have to trust me. You can verify me.

That signal — the willingness to be verified — is one of the most powerful trust signals available in an institutional finance context. More powerful than the relationship. More powerful, even, than the numbers themselves. Because it says: I have nothing to hide, and I know what a model is for.

The node that does not trust

There is something worth pausing on here. The willingness to be verified — the transparent model, the marked inputs, the auditable assumptions — is not a universal professional value. It is a specific cultural expression of a specific model of trust.

Consider its logical endpoint: the Bitcoin node. In a distributed blockchain network, no node trusts any other node. Instead, each node independently verifies every transaction against the mathematical rules of the protocol. Trust is not required because verification is absolute. I trust you because I have calculated that everything you say is true — taken to its extreme, removing human judgment entirely, replacing relationship with cryptographic proof.

That architecture did not emerge from nowhere. It is the product of a specific tradition — one that runs from double-entry bookkeeping through Dutch commercial law through the development of credit ratings and institutional due diligence — in which verification is more reliable than trust, proof is more valuable than relationship, and the willingness to be checked is the highest available signal of credibility.

This is not a universal model. It is primarily a Northern European and Anglo-American model, refined over centuries of commercial practice in cultures where personal networks were insufficient to sustain the scale of trade that was emerging. The solution was not to expand the networks. It was to replace them — with contracts, audits, models, and eventually with cryptographic proofs.

Relational trust cultures — which include much of South Asia, East Asia, the Middle East, and Latin America — developed different and equally sophisticated solutions to the same problem. The solution was relationship. Extended networks of personal obligation, the weight of the introduction, the guarantee of the shared history. Not less reliable than verification. Differently reliable — optimised for different environments, different scales, different kinds of risk.

The financial model sits at the intersection of these two traditions. In the verification tradition, it is the trust instrument — the thing that replaces personal relationship with auditable proof. In the relational tradition, it is the formality — the paperwork that accompanies a commitment already made through the relationship.

When a developer from a relational trust culture presents a model to an investor from a verification culture, both understand that a financial model is involved. Neither may understand that they are using the word to mean entirely different things.

The intercultural inconvenient truth

The developer who built the eighty-tab monster was not necessarily trying to hide something. The colleague who produced the pasted-values simplification was not deliberately concealing the absence of formulas. Both, in their own way, were operating from a model of professional relationships in which the model itself is not the primary trust instrument.

In a relational trust culture, a complex model signals effort and seriousness. All those tabs, all those cells — someone worked very hard on this. The complexity is not concealment. It is evidence of commitment. In an institutional finance culture, the same complexity reads completely differently.

The seasoned investor who told me that it is almost standard — models that try to hide the real costs and risks somewhere in tab 48 — was not describing a conspiracy. They were describing a pattern. A pattern produced by developers who understood the relationship as the guarantee and the model as the formality, presenting to investors who understood the model as the guarantee and the relationship as secondary.

Two models of trust. One spreadsheet. Thoroughly different readings.

The opaque model — whether through shock and awe complexity or through the absence of formulas in a simplified version — maximises the information asymmetry rather than reducing it. The principal cannot verify. They can only trust. And in institutional finance, being asked to trust without the means to verify is not a neutral experience.

Two civilisational traditions of trust, meeting in a spreadsheet. The verification tradition asks: can I calculate that what you say is true? The relational tradition asks: do I know and trust the person saying it?

In institutional finance, the first question is always asked first.

Mark your inputs. Show your formulas. Make the model breathe.

Not because the rules require it. Because the model is already talking. The only question is what it is saying.

Culture & Capital publishes at the intersection of commercial practice and cultural intelligence. The Road Between has been in enough of these rooms to know that the model is never just a model.

Subscribe at culturecapital.media