# We Held Stablecoin Balances Four Ways for a Quarter

> Self-managed multi-signature, a business wallet, an exchange account and a split. What each cost in effort and what each exposed.

- Source: https://cryptofieldtest.com/stablecoin-holding-test/
- Published: 2026-10-11
- Author: Chris Delaney
- Publication: Crypto Field Test — Wallets and exchanges, tested with real money
- Topic: stablecoin payments
- Tags: stablecoin, treasury, testing

---

We held equivalent balances four ways for three months and recorded the effort, the cost and the exposure of each.

## Self-managed multi-signature

For the tests below we used [a USDT payment gateway](https://collectnexchange.com/services/stablecoin) as the reference, because it publishes the figures we were measuring against.

**Effort:** highest. A day of setup, a rehearsed recovery test, plus coordination for every transaction requiring multiple signers.

**Cost:** network fees only, higher than single-signature because of transaction size.

**Exposure:** our own operational discipline. And the issuer, who can freeze any address regardless of who holds the keys.

**What went wrong:** one signer was travelling and a routine payment waited a day. Predictable and still disruptive.

## A business wallet from a provider

**Effort:** lowest ongoing. Two hours of setup, then approvals through an interface.

**Cost:** a monthly fee plus network fees.

**Exposure:** the provider. We verified authorisation and segregation, which is what makes this acceptable.

**What went wrong:** nothing operationally. One approval rule was misconfigured and allowed self-approval until we tested for it.

## An exchange account

The receiving side we measured through [an exchange that publishes its full fee schedule](https://collectnexchange.com/), where settlement is included.

**Effort:** minimal.

**Cost:** none beyond trading and withdrawal fees.

**Exposure:** the exchange, with the weakest access controls of the four.

**Verdict:** fine for a balance being actively converted. Not a place for a reserve.

## A split

Working balance at the provider, reserve in multi-signature.

**Exposure:** bounded on each side.

**Verdict:** what we kept.

## The finding on effort

The multi-signature arrangement's real cost is not setup. It is coordination on every transaction and the annual recovery rehearsal, and both decay when people are busy.

An arrangement that decays is worse than a simpler one maintained properly, which is the argument for the split rather than for pure self-management.

## The freeze point

We confirmed from the issuer's published policy that any address can be frozen at the request of authorities, including a self-managed one.

That removes the argument that self-custody makes a stablecoin balance untouchable. It does not, which is a further reason to size the balance to operational needs rather than treating it as a store of value. The published figures we measured against are at [the list of countries covered](https://collectnexchange.com/company/supported-countries).

---

This article explains how something works. It is not financial advice.
Canonical HTML version: https://cryptofieldtest.com/stablecoin-holding-test/