On registers

There are two ways to reconnect capital with its owner.

The first is the search. Someone, usually the owner or a relative, starts from a name and goes looking. They write to an old employer, phone a fund, fill in a form, wait. If they are lucky, the record turns up. If not, they try another country, another fund, another year. Every search starts from zero.

The second is the register. Owners and holders both leave a mark in one place, once, and the match happens when the two marks meet. Nobody has to know where to look. The register remembers for them.

Most of the world still runs on the first model. It is worth asking why the second keeps winning wherever it is tried.

A search is a one-off transaction. It costs effort every time, and the effort falls on the person least equipped to spend it: the owner, often years later, often abroad, often without the paperwork. When the search fails, nothing is learned. The next person starts from the same blank page.

A register is infrastructure. The cost is paid once, at entry, and every later lookup is cheap. Each new entry makes the register more useful to everyone else. A search is linear. A register compounds.

The idea is not new. Some countries already run central registers for exactly this problem, and they work.

Switzerland is the clearest example. Second-pillar workplace pension money that has been left behind is reported to a central office, the Zentralstelle 2. Säule, and anyone can ask it, free of charge and from abroad, whether vested benefits are held anywhere in the country in their name. One request replaces dozens of letters to individual funds. PensionHunter's guide to the Swiss central vested-benefits search explains how it works for people who have left.

Israel offers another. Its official Har HaKesef search covers inactive financial accounts, including pension and provident accounts, and it does not require a national digital identity to use. An identity number and the issue date on the physical card are enough. That single design choice opens the door to people who have moved away. The details are in PensionHunter's page on finding an Israeli pension from abroad.

These registers share three properties. They are central, so the owner does not need to know which institution holds the record. They are free, so cost does not decide who looks. And they are national, which is also their limit.

A national register can only see its own country. The nurse who worked in four countries needs four registers, if all four countries have one, and most do not. Nothing connects them. The person who moves is, once again, the person the infrastructure was not designed for.

The obvious next step is a register that sits across borders and works from both sides.

From the owner's side, it is a forwarding address for capital. A person who is about to leave a country, or who left years ago, registers once: the countries they worked in, the names they used, a way to reach them. They do not need to know which fund holds what.

From the holder's side, it is a way to find former members. A fund trying to reach people it has lost can register too, and check whether someone it is looking for has left a forwarding address.

The match happens only when both sides agree. Nothing is shared without consent. That is not a courtesy. It is the condition that makes the whole thing trustworthy enough to use.

This is the shape of GPFR, the Global Pension Finding Register, which PensionHunter operates and which is free for people and free for funds. It is early. Its value depends entirely on both sides joining, which is exactly why it has to cost nothing to join.

The register model also changes what a lost record is. Under the search model, a record is lost until someone happens to look in the right place. Under the register model, a record is simply waiting for its other half. The difference sounds small. Over time it is the difference between an archive and a network.

Machines matter here in a specific way. They make it cheap to normalise messy entries: transliterated names, old employer names, former surnames, addresses in different formats. A register that machines can read is one where near-matches are found and then confirmed, rather than missed. This is the same argument as On proof of ownership: finding is becoming cheap, and the value moves to confirming.

The register is also where independence matters most. A register that sold its entries, or steered people to paid advice, would empty quickly. We always point people to the free government tracing route first, and we take nothing from financial advisers. A cross-border register should sit next to the national registers, not in front of them.

The summary, for anyone building infrastructure for stranded capital:

• A search starts from zero every time. A register compounds.
• National registers work, and stop at the border.
• A two-sided register lets owners leave a forwarding address and holders find former members.
• Consent from both sides is what makes it usable.
• Free entry on both sides is not generosity. It is the design.