On September 17, 2026, the publication Banking 4.0 analysed how Italy defends against identity fraud. According to its account, the country relies on four separate layers. SIMoITel covers telecom payment defaults. SIFRID, run through Experian Italia, carries suspicious-application signals. SCIPAFI checks identity documents against public records. CAMARA-based network APIs (SIM Swap, Number Verification and KYC Match) are offered by operators including TIM and Wind Tre. Banking 4.0 argues for purpose-specific trust sources rather than one central identity database. This is analysis of an existing system, not a new regulatory act. It still deserves attention, because many governments are debating whether to tie SIMs to a single national identity record.
The case for a central identity record
The best argument for centralisation is simplicity. One authoritative database gives every bank, telecom and public agency the same answer about who someone is. It leaves fewer gaps for fraudsters to exploit between systems. It also makes accountability clearer, because one body is responsible for accuracy. Regulators who favour tighter SIM-to-identity binding are responding to a real problem. Scam losses in Italy rose from €114 million in 2022 to €137 million in 2023, according to figures in a GSMA Open Gateway announcement republished by MWC Barcelona. Wanting a stronger anchor for identity is a reasonable response to losses of that size.
What Italy built instead
Italy's approach splits the job by purpose. SCIPAFI is the public-sector layer. It was created by Title V-bis (Articles 30-bis to 30-octies) of Legislative Decree 141/2010, which set up a public system for administrative fraud prevention in consumer credit, with specific reference to identity theft. The Ministry of Economy and Finance owns it and Consap manages it. Lenders and insurers use it to check that the data in an application matches public records.
The verification is narrow. It does not issue an identity or profile anyone. It answers specific questions: does this tax code exist, is this document reported stolen, is this person alive. Il Sole 24 Ore, reporting on the ministry's annual fraud report, gave the 2025 results. More than 205,000 tax codes were non-existent (0.64% of those checked). More than 210,000 documents were flagged as stolen or lost (1.12%). More than 23,000 checked names belonged to deceased persons (0.10%). Cases involving deceased persons' identities fell by roughly half against 2024. The report credits integration with the national resident-population register. The system has handled more than 189 million queries since 2015.
Those rates matter. A hit rate of about 1% on stolen documents is high enough to justify the checks and low enough to show that most queries touch honest customers. A design that serves that majority should collect as little about them as it can.
The telecom layer: ask the network, don't copy the register
The CAMARA network APIs apply the same principle to mobile identity. SIM Swap tells a bank whether a number's SIM was recently replaced, which is a common sign of account takeover. Number Verification confirms the device holding a number without an SMS one-time code. KYC Match checks whether details a customer supplied match the operator's records, and it does so without handing over the underlying data. GSMA's Open Gateway initiative reports that TIM and Wind Tre have launched their first CAMARA APIs. The identity and anti-fraud APIs are the stated priority.
The output is a yes/no signal, not a record. A bank learns that a SIM swap happened in the last few days. It does not learn who the customer calls or where they travel. That separation is what a central SIM-to-ID database would erase. If every SIM were bound to a national identity record, one breach, one misconfigured access grant or one change of political mood would expose phone use, financial activity and civil identity together.
Why this fits the law as well as good engineering
The GDPR requires personal data to be collected for specified, explicit and legitimate purposes and not further processed in an incompatible way (Article 5(1)(b)). It also requires data to be limited to what is necessary (Article 5(1)(c)). Purpose-specific trust sources follow these rules by design. Italy's privacy regulator applied the same logic to SCIPAFI early. In a November 2014 opinion, reported by Diritto.it, the Garante allowed the system on condition that participants use only relevant, non-excessive data for specified purposes, retain it no longer than necessary, and restrict access to authorised participants. The ministry's 2025 draft regulation to replace the 2014 rules, which went to consultation with comments due August 7, 2025, reportedly adds GDPR-aligned protections, anti-money-laundering integration and lower query fees.
Limits of the model
The layered approach has costs. Fraudsters can probe the gaps between systems. Coverage is uneven, since some operators are further along than others. The checks serve credit and payments, not every digital service. Banking 4.0's piece also leaves open whether private signals such as SIMoITel and SIFRID have clear rules on accuracy and redress when a customer is wrongly flagged. Those are governance questions to fix. They are not arguments for abandoning the architecture.
What policymakers elsewhere should take from it
First, match each check to a defined risk. Document fraud, SIM takeover and payment defaults need different evidence, so they should not share one source. Second, prefer answers over records. A confirmation or a yes/no signal gives a relying party what it needs and leaves the data where it is. Third, publish the hit rates. Italy's reporting on SCIPAFI lets outsiders judge proportionality. A mandate to bind every SIM to a government ID should have to show that its benefits exceed what these narrower tools already deliver.
Italy's example does not prove that no central identity layer is ever justified. It shows that fraud can be reduced a good deal without one, and that avoiding it keeps the open, low-friction mobile ecosystem that digital services depend on.