Lease or Moat · PNDK Solutions Ltd

Moat or lease? How the protocol tells them apart

What is an economic moat? An economic moat is a durable competitive advantage that protects a company's returns from competitors; Warren Buffett popularised the term. Lease or Moat adds one test: is the advantage owned, or does it run through a constraint another party controls and could withdraw?

A company can grow fast on a position it does not own. The Lease-or-Moat protocol separates velocity — how fast it grows — from capture — whether it keeps the economics. Two gates, five questions.

Gate 1 — Is there a constraint at all?

Before ownership matters, something has to be hard to reproduce. Two tests locate it.

1A · Substitution

Can a competitor or customer get most of what the company provides by routing around it? If most of the value is available elsewhere, there is no constraint to own — only a head start.

1B · Withdrawal

If the decisive dependency were withdrawn, would the position disappear, or would the company and its rivals simply pay more? This half carries the weight. Every company depends on something; the question is whether withdrawal removes the specific constraint that keeps rivals out, or only raises a bill everyone pays. A commodity input raises the bill. A leased constraint removes the position.

Gate 2 — Who captures what the constraint produces?

Once a constraint exists, ownership becomes the decisive variable.

• Ownership. Does the company materially control the constraint, or does an upstream party hold it? A constraint the company does not own is not a moat — it is a lease.
• Rent capture. Who keeps the economics the constraint produces? Control and capture usually move together; where they diverge, that gap is the story.
• Flashlight. Who upstream could turn the company's product into a feature — and is that capability dark (no one positioned to ship it), latent (someone could, but hasn't), or lit (someone already ships it at scale) on the decisive object?
• Transition. If the company is on a path to ownership, name the object that must move from rented to controlled, the evidence that would prove it, and what would reverse the reading.

The three readings

• YES — owned closure. The company controls the decisive constraint and captures its rent. Reported with its scope: most moats are bounded, not total.
• MAYBE — conditional. A credible path to ownership, but the threshold isn't crossed. A build order, not a valuation.
• NO — lease. The constraint is controlled upstream, or the position is routable. The rent leaves with the thing holding it up.

Where it applies

The protocol is not a test for software or AI companies. It applies to any business whose position runs through something that someone else could withdraw, reprice or copy. The question is always the same: what would a rival have to reproduce to take the position, and who controls it? The paper's annex reads companies across sectors and regions:

• A network of participants. M-Pesa's customer–agent–merchant ecosystem reads as a bounded YES in Kenya and as a MAYBE in Ethiopia, where the same network is still forming. The protocol reads the position, not the name.
• Production, craft and brand. Hermès reads YES on a non-technology closure made up of several objects together: production capacity, craft, distribution and brand.
• An integrated technical system. ASML buys scarce inputs and still reads YES. The scarcest input is not automatically the constraint.
• A rail owned by others. Paystack owns the merchant interface but runs on bank and card rails, so it reads NO: owning the interface while renting the decisive constraint.
• A grant, a site or a licence. A regulatory grant, certified storage, a port concession or a licensed dataset is rented if the grantor can end it, however scarce it is. Tenure decides.

External dependence alone does not make a lease. Every business depends on something. The protocol asks whether the dependence is the object that keeps rivals out.

Why evidence matters

The protocol asks the right questions, but an answer is only as good as the observable fact behind it. A reading should distinguish an observed fact from a warranted claim from a structural conclusion — and say plainly where it is asserting rather than showing. The instrument records the fact behind each load-bearing answer and prints it into the reading, so the verdict never says more than its evidence warrants.