Credit Data for Economic and Population Research

Pick the right data points to support your research projects.

TransUnion regularly works with researchers at universities, non-profits and government agencies to deliver de-identified credit data to enable their economic and population research. This kind of data opens significant opportunities for population sampling, understanding consumer patterns of decision-making, and evaluating financial outcomes across different moments in time. However, working with credit data and navigating the breadth and depth of information can be challenging.

Where does credit data come from?

Credit data is reported to TransUnion from a wide network of data furnishers. TransUnion’s database of credit histories (covering more than 260 million US adults) is supported by a network of over 85,000 national, regional and local furnishers — which includes finance companies, banks, credit unions, retailers, student loan providers, debt buyers, and more. We require all data furnishers to provide data that’s complete and accurate in accordance with the FCRA and all applicable state credit-reporting laws, and in adherence to the Metro 2® Reporting Guidelines maintained by the Credit Data Industry Association.

The atomic unit of credit data is the account, and accounts can be classified as tradelines or collections.

• A tradeline captures an ongoing, historical and current record of a particular debt obligation issued by a lender. This includes things like car loans, mortgages, credit cards and student loans.
• ·A collection is the result of some unpaid financial obligation. A collection can be associated with a tradeline (e.g., defaulting on a credit card), but others can be reported in circumstances where an entity is owed payments by a consumer outside of a lending context (e.g., an unpaid parking ticket).

Consumers’ credit reports are a compilation of the different accounts associated with that consumer — as reported by these furnishers at a particular moment in time. TransUnion’s consumer reporting database also includes personally identifiable information (PII) — which is never shared for use in research.

What account information is reported by data furnishers?

Data furnishers provide various levels of credit account information, including financial details, descriptors and remark codes.

Account financials

Most consumers are familiar with the basic fields shown on their credit reports which capture the financial details of an account at a given moment in time (usually the end of the account’s billing cycle) and includes data points like:

• Credit limit
• Current balance
• Minimum payment due
• Amount past due
• Account status (current, 30 days past due, etc.)

Account descriptors

Data furnishers also report static properties of each account that do not typically change from month to month, such as when the account was opened or who the lender is. More important for a research audience are three key pieces of data required by the Metro 2 guidelines that provide a flexible framework for capturing the structural details of a given financial obligation:

• Account type: A code describing the kind of obligation the account represents, such as an auto loan or lease, a credit card or student loan.
• Equal Credit Opportunity Act (ECOA) code: A code identifying the consumer’s liability for payment on the account, such as whether they are individually responsible for payments or a co-borrower with joint responsibility.
• Portfolio type: A code providing a high-level classification of the payment expectations on the account, allowing for delineation between revolving accounts, installment accounts, and open accounts where the balance is due upon demand.

Researchers can flexibly use these fields for a view of consumer credit obligations based on project needs. For example, the right combination of values would allow a researcher to sample from consumers who are authorized users on a private label, retailer-specific credit card issued to another consumer without a personal, bank-issued credit card of their own.

Account remark codes

Occasionally, data furnishers will append ‘remarks’ to explain a special condition related to an account. The dates when remark codes are added or removed are also maintained in an account’s history. These remarks cover a variety of circumstances, including:

• Account included in consumer bankruptcy
• Accuracy of account disputed by consumer
• Account affected by a natural disaster
• Account closed due to refinance

This information can aid researchers in identifying consumers who’ve had (and particular points in time when) special circumstances occurred, such as reporting a credit card lost or stolen, or the date an account was first flagged as having payment assured by wage garnishment.

What are credit-derived attributes, algorithms and scores?

In our experience, researchers use depersonalized, consumer-level data rather than account-level data. In addition to maintaining account records, TransUnion calculates consumer-level attributes, algorithms and scores derived from each consumer’s account information on a regular basis. These provide a streamlined way for researchers to operationalize credit data for model development, hypothesis testing and analytic inquiry compared to raw tradeline data.

• Attributes are essentially calculations quantifying what can be directly observed in a consumer’s credit history at a given point in time, such as the balance of all open auto loans or total credit utilization across all revolving accounts.
• Algorithms classify consumers based on observed credit behavior, such as whether a consumer is a “Revolver” or a “Transactor” on their credit cards, or how frequently they’ve made prepayments on personal loans in the past.
• Scores are models designed to predict the likelihood of some outcome occurring over a certain time frame as of a particular moment in time, such as applying for a personal loan within 6 months, repaying an outstanding collection within 1 year or having an account go more than 90 days delinquent in 2 years.

Working with TransUnion to use credit data in economic and population research

We take a consultative approach to working with researchers interested in our data. We’ve worked with experts in a variety of fields of study, including economics, public health, government policy and marketing research. Our team can provide valuable insights and best practices for experimental design, sampling from our national credit database, selecting the right attributes for impact assessments or model building, and more.