How to Read an Exchange's Proof of Reserves (POR)?

What is POR? Why has everyone been talking about it after FTX? What can it actually prove, and what can it not? These three questions are the right entry into Proof of Reserves. POR was once marketed as the “ultimate transparency tool,” but pull its logic apart and you’ll see it’s far from omnipotent — a genuine step forward and a document easy to over-read at the same time.

What POR is: which question is it answering

POR’s purpose is to let an exchange demonstrate, in a way outsiders can verify, that “at a given moment, the on-chain assets we hold are enough to cover all user balances.”

Three limits are baked in:

• At a given moment — a snapshot, not continuous monitoring.
• Enough to cover — assets ≥ liabilities, not exactly equal.
• Outsiders can verify — users can check that their own balance is included.

Before FTX, “are user balances actually backed?” rested on the operator’s word. After FTX, “trust the executives” wasn’t enough, and POR was pushed forward. But understand this clearly: POR addresses a slice of the question at one point in time — not “are you always backed.”

Merkle trees: proving one thing with a tree

The core tool POR uses is the Merkle tree. Simplified, it works like this:

User balances are the leaves of the tree. Each pair of leaves is hashed into a parent node, then pairs of those into the next layer, and so on until a single Merkle root sits at the top. This tree has a key property:

• The smallest change to any leaf changes the root hash all the way up.
• You can prove your own balance is in the tree without revealing other users’, using only the path from your leaf to the root.

When the exchange publishes the root hash, every user can plug their account balance into a verification tool, get the matching path, and compare against the published root — confirming “my share is included in the platform’s stated liability total.” The asset side is shown through on-chain wallet addresses. Compare the two sides and you, in theory, see whether assets cover liabilities.

What it can prove

Honestly listed, POR’s actual capabilities are three.

First, user presence. You can use the Merkle proof to verify your own balance isn’t silently dropped. Moving from “the platform says so” to “I can verify my share is in the list” is a real improvement.

Second, assets ≥ liabilities at a point in time. When an auditor or third party matches the asset-side wallets against the Merkle-aggregated liabilities at one moment, you can confirm coverage at that instant.

Third, on-chain observability. Once reserve addresses are public, anyone can monitor them on-chain, raising the odds that anomalous flows get spotted.

Variations in implementation

Exchanges do POR differently along three axes: whether the liability side uses a Merkle tree, whether the asset side signs to prove control, and whether an independent third party is involved.

The liability side usually uses a Merkle tree. A few platforms only publish “total user balance” without a verifiable path — that’s a number screenshot, far less meaningful than a real Merkle proof. On the asset side, two approaches dominate: publishing reserve addresses for anyone to inspect, or signing messages to prove control so the exchange can’t borrow someone else’s coins for show. Third-party involvement varies — some hire known audit firms, others publish in-house reports. Whether an independent party is in the loop directly affects credibility.

Its limits: what POR cannot prove

This is the most easily ignored and most important part.

Limit one: the time blind spot. POR is a snapshot; what happened between two snapshots is invisible. A platform could borrow assets for a day, pass POR, then move them away — the report still looks “clean.”

Limit two: liability-side fraud. The Merkle tree proves “your balance is in the list,” not “the list equals the real total.” If a platform deliberately omits internal accounts or undisclosed books, a regular user can’t see it.

Limit three: off-chain liabilities. POR only covers on-chain assets and crypto liabilities. Fiat debt, internal obligations, implicit commitments — outside scope.

Limit four: continuity. Most PORs publish quarterly or semi-annually. Far from real-time transparency.

These limits don’t make POR useless; they remind you it is one signal among many, not a disclaimer.

How a regular user can use POR as a reference

In practice, four things to look at. First, does the exchange offer a personally verifiable Merkle proof — can you input your account, get a path, and compare against a root hash? Second, does the asset side publish concrete reserve addresses with signed proof of control? Third, what’s the cadence — quarterly or more frequent — and do reports line up over time? Fourth, is an independent third party in the loop?

More importantly, place POR inside your full exchange selection criteria, not on its own. An exchange with POR but poor compliance and messy custody isn’t safer than a steady-running platform without one. POR is a plus, not a free pass.

POR is a tool, not a disclaimer

Back to the three opening questions. What is POR? A point-in-time “assets ≥ liabilities” proof backed by technical machinery. Why does everyone talk about it after FTX? Because it moves part of what used to rest on words onto verifiable math. What can it prove and not prove? It can confirm your balance is included in the disclosed liability set and that on-chain coverage held at that moment; it can’t prove ongoing solvency, that the liability side wasn’t doctored, or anything off-chain.

At that depth, you won’t think “passed POR” means safe, and you won’t dismiss reports just because they’re technical. POR is far better than nothing and nowhere near enough: it shifts a sliver of trust from words onto math, but the longer road past that point still depends on a black-swan plan, diversification, and disciplined withdrawals. Treat it as one signal among many, not a passport — that’s how to read POR.

This article is educational and does not constitute investment advice. POR only provides point-in-time, on-chain coverage proof; combine multiple signals when evaluating exchange risk.