Financial AtlasTo the mapУкраїнська

SERVICE INFRASTRUCTURE

Credit Information

The system's memory of debts: who paid whom and how — and how that becomes a credit score.

Open on the interactive map →

This page is the text version of one zone of the interactive map.

Its role in the system

The financial system's memory of who repaid debts and how. Credit bureaus and public credit registers collect the payment history of people and companies, and scoring models turn it into a risk estimate. This information decides whether a loan is granted, for how long and at what rate — and whether the supervisor can see a country's overall debt burden.

Functions

  1. Credit histories — bureaus collect data on loans, payments and arrears from banks, consumer lenders, leasing and telecom companies and build a file on the borrower (🇺🇸 Equifax, Experian, TransUnion; 🇩🇪 SCHUFA; 🇮🇹 CRIF).
  2. Public credit registers — central banks and supervisors keep their own loan databases for supervision and risk analysis (🇪🇺 AnaCredit at the ECB — loans to legal entities above 25 thousand euro; 🇮🇪 the Central Bank of Ireland's Central Credit Register).
  3. Scoring — statistical and machine-learning models turn history into a number that estimates the probability of default (🇺🇸 FICO on a scale from 300 to 850, VantageScore; banks' internal models for Basel ratios and IFRS 9 provisions).
  4. Alternative data — utility, rent and telecom payments, open banking data, behavioural signals: they give a chance to people who have no classic credit history yet.
  5. A person's rights over their own data — access to your own file, challenging a wrong entry, the time limits after which negative information disappears.
  6. Identity verification and anti-fraud — the same data is used to detect identity theft, "synthetic" borrowers and forged applications.

What makes it special

  • How it earns, and why a register is not a bureau: a private bureau sells reports and scores to the same lenders that supply it with data — reciprocity applies: you get access to the database only if you report into it yourself, and the bureau exists for the lender, to assess one specific application. A public credit register sells nothing: the central bank maintains it to see a client's total debt across all banks and the risks of the system as a whole. In many countries both operate side by side.
  • Who regulates: 🇺🇸 the FCRA (1970), supervised by the CFPB and the FTC; 🇪🇺 GDPR plus the new Consumer Credit Directive 2023/2225, which applies from 20 November 2026; national data protection authorities handle complaints about bureaus.
  • A score is a forecast, not a judgement of a person: the model computes the probability that a borrower will stop paying, not how "good" the client is, and different bureaus and models produce different numbers for the same person because they hold different data and weight features differently. Hence the core tension: data about a person is collected without that person asking for it, yet an error or a leak hits precisely that person — which is why regulators and courts increasingly demand accuracy of data, explainability of models and a real right to challenge a decision.
  • 🇨🇳 China: a state register at the centre of the system: the backbone is not a private bureau but the Credit Reference Centre of the People's Bank of China, which lenders report into: by August 2022 its files covered roughly 1.15 billion individuals and almost 99 million companies, and a query to it has become a compulsory step in any lending decision. Only two market-based bureaus for individuals hold a central bank licence — Baihang (2018) and Pudao (2020) — and the 2022 Measures for the Administration of the Credit Reporting Business allow only the minimum necessary data to be collected and require the person's consent and their rights to know, object and complain.
  • 🇯🇵 Japan: three bureaus split by type of lender: the Personal Credit Information Center at the Bankers Association (KSC) serves banks, CIC serves instalment credit, cards and mobile operators, and JICC serves consumer finance companies; since March 1987 they have exchanged negative information through the CRIN network, so a missed payment recorded at one bureau is visible to all three.
  • 🇸🇬 Singapore: the bureau works under a central bank licence: the Credit Bureau Act 2016 has been in operation since 31 May 2021 — the Monetary Authority of Singapore licenses credit bureaus and supervises them much like financial institutions, and a person has the right to obtain their report, review it and have an error corrected. Disputes between a customer and a bureau go to FIDReC, the independent dispute resolution centre a licensed bureau must belong to. Bank and finance company data is pooled by the licensed Credit Bureau (Singapore); for licensed moneylenders the Ministry of Law keeps a separate bureau, the MLCB, which they must report into and must check before granting a loan.

Why it matters

When credit information is inaccurate, unavailable or leaked, the system errs in both directions: lenders hand out money blindly and pile up problem debts — or refuse people who are perfectly able to pay. And those with no history at all ("credit invisibles") are shut out of borrowing altogether: on the CFPB's estimate, as of December 2020 that was 2.7% of American adults, with almost another 10% holding records from which no score can be calculated.

🇺🇸 The Equifax data breach, 2017

One of the three largest US credit bureaus failed to install a security update for a known vulnerability, and attackers quietly siphoned off data for months — around 147 million people were affected, including 145.5 million social security numbers. In 2019 Equifax agreed a settlement with the FTC, the CFPB and the states worth at least 575 million dollars, with the fund for affected people able to grow to 425 million. These people were not Equifax customers and had no way of influencing the fact that their data was held there.

🇪🇺 The Court of Justice of the EU in the SCHUFA case, December 2023

The German bureau SCHUFA calculated the score while the formal credit decision was taken by the bank — so for years the bureau argued that it decided nothing about anyone. On 7 December 2023 the Court of Justice of the EU ruled that if the lender in practice always follows the score, then calculating the score is itself an automated decision within the meaning of the GDPR, with a right to an explanation and to challenge it. In February 2025 the court added that the explanation must be comprehensible to a person, and that trade secrecy is not a ground for refusing one.

🇮🇪 Ireland's Central Credit Register, 2017–2019

Before the register existed, an Irish lender could not see how much the same borrower already owed to other lenders — one of the reasons the credit boom before 2008 left the country buried in bad loans. The Credit Reporting Act 2013 gave the Central Bank of Ireland a register: consumer loans began to be reported from mid-2017, business loans from 2018 and hire purchase from 2019, and a lender must check the register for any application of two thousand euro or more. The borrower can obtain their own report free of charge and demand that a wrong entry be corrected.

Links to other zones · 9

  • Banking & Credit InstitutionsBanking & Credit Institutions are both the main suppliers and the main users of Credit Information: they report payment histories daily and request a report and a score before every credit decision.
  • Corporations & Corporate FinanceCorporations & Corporate Finance have their own credit files and business scores in Credit Information; registers reveal the group's total borrowing, and suppliers check them before shipping goods on credit.
  • Debt marketWhen a pool of loans is packaged into Debt market securities, investors and models rely on borrower data from Credit Information, and the quality of that data sets how truthful the risk assessment is.
  • Digital paymentsCredit Information leans more and more on Digital payments data: a record of incoming and recurring payments gives a chance to people with no classic credit file, and sharpens the question of consent and privacy.
  • Households & IndividualsA person's file in Credit Information decides whether Households & Individuals get a mortgage, a card or instalments and on what terms; a wrong entry or a data leak hits the household, so the law grants access and correction rights.
  • Monetary PolicyThe central bank behind Monetary Policy runs its own register inside Credit Information and uses it for financial stability analysis, stress tests and tracking how rate changes reach real borrowers.
  • Ratings & IndicesCredit Information and Ratings & Indices both judge creditworthiness, but from opposite ends: agencies assign ratings to issuers and bond issues, while credit bureaus score millions of individuals and small companies.
  • Regulation & SupervisionRegulation & Supervision licenses and inspects the bureaus of Credit Information, obliges lenders to check affordability before lending, and defends the consumer's right to accurate data.
  • Technology & InfrastructureCredit Information lives on Technology & Infrastructure: data pipelines, machine learning, open banking APIs, anti-fraud platforms and the cybersecurity without which a credit history database becomes a target.

Careers in this zone · 8

Scoring analyst / model developer

Builds the models that use credit history and other data to compute the probability that a borrower will not repay. Prepares the data, selects features, tests the model against historical defaults and watches that it does not lose accuracy over time. Separately, proves that the model does not discriminate on prohibited grounds.

What you need to know

  • Statistics and machine learning: logistic regression, gradient boosting
  • Model quality metrics: Gini, AUC, KS, population stability (PSI)
  • The credit risk parameters PD, LGD, EAD and Basel requirements
  • IFRS 9 — expected credit losses
  • Fair lending and data protection rules (🇺🇸 FCRA and ECOA, 🇪🇺 GDPR)

Skills

  • Python / R, SQL
  • Working with large and messy data
  • Explaining the model's logic to the business and the regulator
  • Careful model documentation

Certifications: Usually not required; FRM, CFA and a statistics degree are valued

Scoring model validator

Independently of the developers, checks whether the model actually works: do the forecasts hold on new data, are the features stable, is information from the future leaking into the model. Writes the opinion that the risk committee and the supervisor read, and has the power to keep a model out of production.

What you need to know

  • Model validation and backtesting
  • Regulatory requirements for model management (🇪🇺 EBA guidelines, 🇺🇸 SR 11-7)
  • Statistics and econometrics
  • Credit risk and portfolio structure

Skills

  • Python / R
  • Critical review of someone else's work
  • A clear written opinion
  • Holding an independent position

Certifications: FRM (GARP), PRM; CFA as a plus

Credit bureau data quality analyst

Makes sure the data sent in by banks, consumer lenders and leasing companies lands in the database correctly: matched to the right person, without duplicates or stale records. Resolves discrepancies with data providers and keeps track of retention periods.

What you need to know

  • The structure of a credit history and bureau reporting formats
  • Rules for matching a person to a record (identifiers, matching)
  • Retention and deletion periods for negative records
  • The basics of credit products

Skills

  • SQL and work with large datasets
  • Data quality control
  • Communication with data providers
  • Attention to detail

Certifications: Usually not required — hands-on experience with data counts

Regulatory credit reporting specialist

Prepares a bank's submissions to public credit registers (🇪🇺 AnaCredit, 🇮🇪 the Central Credit Register): gathers data on every loan across hundreds of prescribed attributes, reconciles it with the books and fixes deviations from the regulator's rules before the filing deadline.

What you need to know

  • Credit register requirements: AnaCredit and national registers
  • Bank accounting and credit products
  • Regulatory reporting and its calendar
  • Data quality management

Skills

  • SQL and work with reporting systems
  • Working to a hard filing schedule
  • Interpreting regulatory instructions
  • Investigating discrepancies in data

Certifications: Usually not required — experience in bank reporting counts

Data protection specialist (privacy / DPO)

Makes sure the collection and use of credit data has a lawful basis, and that a person can find out what is held about them and correct an error. Assesses the risks of new data sources and new models, and prepares responses to requests from people and from the regulator.

What you need to know

  • 🇪🇺 GDPR, in particular Article 22 on automated decisions
  • 🇺🇸 FCRA and consumer rights over credit reports
  • Data protection impact assessment (DPIA)
  • The life cycle of credit data: sources, storage, deletion

Skills

  • Legal analysis
  • Negotiating with the business and data teams
  • Explaining people's rights in plain language
  • Maintaining the record of processing activities

Certifications: CIPP/E and CIPM (IAPP); a law degree

Credit fraud analyst

Looks for applications built on identity theft, invented "synthetic" borrowers or organised groups. Builds detection rules and models, works through suspicious cases by hand and feeds the conclusions back into scoring so the system learns.

What you need to know

  • Credit fraud typologies, in particular synthetic identities
  • Methods of verifying identity and documents
  • Behavioural and device analytics
  • The credit process from application to disbursement

Skills

  • SQL / Python
  • Investigation and linking applications to each other
  • Working with anti-fraud platforms
  • Deciding on incomplete data and in a short time

Certifications: CFE (ACFE) — for fraud investigation

Credit data product manager

Creates the bureau's products: credit reports, scores, monitoring services for lenders and for consumers themselves. Decides which data may lawfully be used, how to present the result to the lender and how to explain it to the person it concerns.

What you need to know

  • The credit data market and lenders' needs
  • Limits on data use and consents
  • The unit economics of information products
  • The basics of scoring

Skills

  • Product analytics, SQL
  • Customer research
  • Prioritisation
  • Working with lawyers and model developers

Certifications: Not required

Disputes and credit history correction specialist

Handles cases from people who disagree with an entry in their credit history or with a refusal of credit: asks the lender for evidence, corrects or deletes the record and explains the decision to the person. Trust in the whole credit data system rests on how well this job is done.

What you need to know

  • Consumer rights in credit information (🇺🇸 FCRA, 🇪🇺 GDPR)
  • Dispute procedures and time limits
  • Typical sources of errors in credit data
  • The basics of credit agreements

Skills

  • Handling complaints and difficult conversations
  • Analysis of documents and evidence
  • Clear writing in plain language
  • Meeting procedural deadlines

Certifications: Usually not required — legal or banking training is valued

All zones of the Atlas