The observatory

The record nobody keeps.

Insolvency is public in every country that has it. Almost nowhere is it comparable. We are researching whether that missing layer can be built from official public sources.

Research only. Everything below is a dated snapshot produced by connector prototypes from heterogeneous public-source records. A record is a publication, not a unique proceeding, and counts may include duplicates and gaps. Nothing here is a coverage, completeness, freshness, real-time or market-availability claim, no jurisdiction is served commercially, and no case data from any administrator appears in this dataset.

public-source records normalised onto one research schema — heterogeneous records, not unique proceedings
records carrying a closing date
records naming an appointed administrator
jurisdictions in this snapshot, of nine research maps
“Despite considerable efforts to produce data on the outcome of insolvency proceedings, little data is available to date.”
European Commission, impact assessment SWD(2022) 395

That is the regulator describing its own continent. The European register of insolvencies connects twenty-two of twenty-seven member states, stores nothing centrally, and offers no history and no interface for machines. Italy and Greece are not in it at all. Germany deletes each case within six months of its closure, so a German time series cannot be reconstructed — only accumulated, starting now.

Coverage

Nine jurisdictions are research maps. Three have a connector prototype that has produced this snapshot; the others have not been attempted yet. Coverage per country is partial and undated beyond the snapshot above, and the figures are not comparable across jurisdictions without jurisdiction-specific validation.

A first finding

How long a French insolvency actually takes.

Duration is the number the Commission says does not exist. Here it is for France, derived from paired opening and closing publications in the BODACC. It is the kind of measurement that only exists once somebody normalises the whole register.

Observations are debtor entities with both an opening and a closing publication in the dataset. Duration is measured between publications, not between court decisions.

The full dataset is not public yet.

If you are an administrator, a fund, a law firm or a researcher and you want access, tell us what you would use it for.

Request access
Early access

Help shape the workflow for your role.

We work directly with administrators, courts, creditors and debtor teams to shape each module around how the case is actually run.

Request a research conversation

This website has no sign-up form and collects no name, email or role. Clicking below opens your own email client — nothing is submitted to or stored by this website.

Data boundary: Trustee OS separates public-source records from private case data. The architecture is designed so private case data would not be reused across clients, cases, public research or cross-client training, and no client data is accessible from this website or from the demonstration prototype. See Security and Privacy.

Status: preview build. Trustee OS is not connected to any court, demonstration content is synthetic, and selecting a country configures the product experience rather than guaranteeing complete local data or court integrations.

Built by Gabriel Casella.