On January 6, 2016, the Federal Trade Commission (FTC) released a report on the growing use of “big data” which discusses potential benefits and risks to big data use and offers practical and legal considerations for businesses.  The report, Big Data: A Tool for Inclusion or Exclusion? Understanding the Issues, focuses on the potential impact of big data on low-income and underserved populations and addresses a host of considerations related to how such data is used.  The report follows the FTC’s public workshop, “Big Data: A Tool for Inclusion or Exclusion,” on September 15, 2014.  Our advisory discussing that workshop is available here. 

The report begins by describing big data and the typical life cycle phases involved, and then explores possible benefits and risks to big data use.  The report concludes by identifying potentially applicable laws and offering legal and compliance considerations for businesses using big data.  Notably, the report expressly states that it was “not intended to identify legal or policy gaps,” but rather guide companies on existing laws that may apply to big data practices.¹  That focus suggests that the Commission believes that existing laws provide for meaningful regulation of big data and that the FTC plans to use existing authority to address big data practices it deems problematic.

Commissioner Ohlhausen issued a separate statement indicating that she supported today’s report and acknowledged that concerns about potential effects of inaccurate data are legitimate, but noted that businesses have strong incentives to ensure accuracy and that free-market competition may inherently resolve the identified issues. 

What is Big Data and How is it Used?

The report explains that the term “big data” does not have a singular definition, but refers to “a confluence of factors, including the nearly ubiquitous collection of consumer data from a variety of sources, the plummeting cost of data storage, and powerful new capabilities to analyze data to draw connections and make inferences and predictions.”²  Big data is often characterized by reference to the “three Vs”: volume (the sheer amount of data that can now be collected and analyzed); velocity (the speed at which industry can collect it); and variety (the breadth and diversity of data). 

In synthesizing discussions from the workshop and the sixty-five public comments submitted, the FTC divided the lifecycle of big data into four phases: (1) collection; (2) compilation and consolidation; (3) data mining and analytics; and (4) use.  The Commission’s May 2014 report, Data Brokers: A Call for Transparency and Accountability, focused on the first three phases of the lifecycle, while this report focuses on uses of big data and potential benefits and risks to use for underserved populations.

Potential Benefits of Big Data

In terms of benefits, the Commission acknowledged that using algorithms to identify patterns could facilitate efficiently matching products and services to consumers.  The report identifies a number of ways that big data is already being used to the benefit of low-income and underserved communities, including by:

• Increasing educational attainment by analyzing big data to identify students for advanced classes that might not otherwise be eligible or by examining trends that might otherwise be left unexamined but result in a better understanding of educational effects. 

• Providing access to credit by using big data to provide alternative credit scores based on non-traditional data such as educational history, professional licensure data, and personal property ownership data for populations that previously were deemed unscorable. 

• Providing tailored health care solutions to individuals and underserved communities by using big data sets to consider individual variability in genes, environment, and lifestyle in developing disease prevention and treatment plans.

• Increasing equal access to employment by using big data to consider shortcomings and potential changes to employer hiring practices.

Potential Risks of Big Data

The Commission noted that researchers and others have expressed concern that big data analytics could be used to make predictions that disproportionately impact certain populations, such as by excluding certain populations from targeted service offerings based on inaccurate predictions.  While big data may be great at showing correlations, large data sets may identify spurious correlations that lack any element of causation, according to the report.  Reliance on such trends may result in detrimental effects on low-income and underserved populations, such as by:

• Denying opportunities based on the actions of others, for example, by lowering a customer’s credit limit based on analysis of other customers.

• Reinforcing existing disparities by targeting ads for financial products such that low-income consumers may never receive ads for better offers.

• Exposing sensitive information such as sexual orientation, ethnic origin, or alcohol, drug and cigarette use.

• Creating new mechanisms for exclusion, for instance, by using identified trends (e.g., individuals who install non-default web browsers are better employees) to support hiring decisions. 

Current Law Governing Big Data and Related Considerations

The report acknowledges that companies will inevitably continue to use big data as a reality of today’s marketplace, but emphasizes that companies should ensure they have an adequate understanding of laws governing big data before doing so and adapt practices to reflect considerations under those laws.  While the Commission acknowledges that current laws may not address every instance of potential misuse of big data, it notes that the report is intended to provide an overview of the existing framework and suggests that the Commission believes it can exercise significant oversight under the current regime. 

The report addresses three statutes specifically: (1) the Fair Credit Reporting Act; (2) equal opportunity laws, including the Equal Credit Opportunity Act, and (3) the FTC Act.

The Fair Credit Reporting Act

The Fair Credit Reporting Act (FCRA) requires consumer reporting agencies (CRAs) that compile and sell consumer reports to implement certain policies and procedures to ensure the safety and accuracy of reports, provide access to consumers’ own information, and follow reasonable procedures to correct identified errors.  “Consumer report” is defined broadly and generally includes reports bearing on a consumer’s credit, character, general reputation, personal characteristics if that report is to be used for a specified purpose such as credit, insurance or employment.³

Consumer reports are only allowed to be provided when the receiving entity has a permissible purpose, which may be the consumer’s own written authorization or for credit, employment, insurance or housing determinations.  Moreover, users of consumer reports are required to take certain steps and provide disclosures when they take adverse action as a result of information contained in a consumer report.

While consumers typically think of conventional credit bureaus or background screening services as CRAs, the definition under FCRA is substantially broader and includes any entity that regularly furnishes consumer reports to third parties for fees.⁴  This means that data brokers and other entities providing reports with consumer information may constitute CRAs under FCRA, even if the provided reports do not contain conventional information typically used in consumer reports.  Indeed, the FTC’s report highlights enforcement actions that the Commission took with respect to online data brokers that were supplying consumer data for purposes that rendered the information subject to FCRA but that failed to comply with FCRA’s requirements.  

For example, in United States v. Spokeo, the Commission brought charges against a data broker for merging online and offline data to create detailed personal profiles that were used by human resource departments for hiring decisions.⁵  Because such reports were used for employment purposes, they constituted consumer reports under FCRA but Spokeo had not complied with the requirements for providing such reports under FCRA.  The report also notes that the FTC has brought actions against users of consumer reports who have failed to comply with requirements under FCRA to use the reports.  In 2013, the Commission brought charges against Time Warner Cable as a consumer report user on the grounds that, according to the complaint, it should have provided certain consumers with a risk-based pricing notice under FCRA and the Risk-Based Pricing Rule.⁶

While FCRA provides a potentially powerful tool to address a company’s use of third party data, it does not generally apply to companies when they use data derived from their own relationship with the customer.   However, the Commission posits in the report that an unaffiliated company’s aggregation and evaluation of a company’s own data would likely make the unaffiliated company a CRA and the other company a user of consumer reports subject to FCRA.    

The Commission also articulated a novel position⁷ that a report could still constitute a “consumer report” even if it does not identify a specific consumer, provided it is crafted for eligibility purposes with reference to a particular consumer or set of particular consumers.  The report distinguished between generating an analysis of a group that shares characteristics with the consumer or consumers being evaluated (potentially a consumer report) from pulling an existing analysis of a characteristic that was not created based on a particular consumer or consumers (likely not a consumer report). 

Ultimately, a company’s obligations under FCRA relative to big data practices will be highly fact specific and depend on the scope and specificity of the report and how the report is used, amongst other factors.

Considerations for Legal Compliance

The report offers the following considerations for companies compiling and/or using big data relative to FCRA:

• CRAs engaged in big data analytics should review the accuracy and privacy provisions of the FCRA, which include requirements to:

(1) have reasonable procedures in place to ensure the maximum possible accuracy of the information you provide,
(2) provide notices to users of your reports,
(3) allow consumers to access information you have about them, and
(4) allow consumers to correct inaccuracies.

• Users of consumer reports should review the provisions applicable to users of consumer reports under the FCRA, such as the “permissible purpose” provisions and adverse notice requirements.

Federal Equal Credit and Employment Opportunity Laws

Companies using big data analytics should also review and familiarize themselves with the federal equal credit and employment opportunity laws, which prohibit discrimination based on protected characteristics. These include the Equal Credit Opportunity Act (“ECOA”), Title VII of the Civil Rights Act of 1964, the Americans with Disabilities Act, the Age Discrimination in Employment Act, the Fair Housing Act, and the Genetic Information Nondiscrimination Act.

Disparate Impact or Disparate Treatment

These sector-specific anti-discrimination laws require proof of either disparate impact or disparate treatment, both of which are theories the FTC has advanced successfully in its ECOA-related enforcement actions.  Disparate treatment occurs when an entity treats an applicant differently based on a protected characteristic such as race or national origin.  Among other things, the ECOA prohibits discrimination based on whether an applicant receives public assistance.  The report cites to several disparate treatment theory enforcement actions where the FTC alleged that lenders excluded public assistance income in deciding whether to extend credit to a consumer in violation of the ECOA.

The report also highlights an FTC ECOA action involving a disparate impact analysis. A disparate impact analysis involves a facially neutral policy that has a disproportionate adverse impact on a protected class.  The policy may violate the ECOA if it has a disparate impact unless it furthers a legitimate business need that cannot reasonably be achieved by means that are less disparate in their impact.  In FTC v. Golden Empire Mortgage, Inc., the FTC alleged that the mortgage lender charged Latino mortgage loan applicants higher prices than non-Latino white applicants and failed to appropriately monitor loan officers and branch managers.⁸  In Golden Empire, there was no legitimate business need justifying this pricing disparity.

Advertising and Anti-Discrimination Laws

Using big data for targeted advertising may also implicate sector specific anti-discrimination laws. Indeed, the FTC’s report explored whether a credit product advertisement targeted to a specific community would implicate equal opportunity laws.  Assuming the offer is open to all to apply and there is no disparate treatment or unjustified disparate impact in subsequent lending, there is likely no equal opportunity law violation.  However, the Commission cautioned that companies should review Regulation B, the implementing regulation for ECOA, in tandem with a targeted advertising campaign. Regulation B has certain record keeping requirements for prescreened solicitations and prohibits creditors from making statements to applicants that would discourage, on a prohibited basis, a reasonable person from pursuing an application.

Similar to a FCRA analysis, whether an employment or credit practice is unlawful will be fact and case specific.  Nevertheless, it would be prudent to review credit or employment practices that rely on big data to ensure they do not discriminate on the basis of protected characteristics in violation of these laws.

Considerations for Legal Compliance

The report offers the following considerations for entities using big data relative to compliance with equal credit and equal opportunity laws:

• Creditors using big data analytics in a credit transaction should review the adverse action requirements under ECOA and requirements related to requests for information and record retention.
• Big data analytics decision-makers that approve credit, housing, or employment applications should consider whether they are treating people differently based on a prohibited basis, such as race or national origin. 
• They should likewise consider whether their policies, practices, or decisions have an adverse effect or impact on a member of a protected class, and if they do, whether that impact is justified by a legitimate business need that cannot reasonably be achieved by means that are less disparate in their impact.

Section 5 of the FTC Act

Companies should also consider whether, in using big data analytics, they are violating any material promises to consumers or making omissions of material facts that are likely to mislead consumers.  Such acts or practices may violate Section 5 of the FTC Act, which prohibits unfair or deceptive acts or practices.  The Commission’s report offers several examples of enforcement actions where it initiated actions against companies for violating material promises to consumers under its deception authority.  For example, the FTC brought an action against CompuCredit for failing to disclose that the company used a behavioral scoring model to reduce credit lines when consumers used their cards for cash advances or certain types of transactions when the company touted the capacity for consumers to use cards for those purposes.⁹

The report also points to the applicability of FTC’s unfairness authority to the misuse of big data.  The elements of an unfairness claim under the FTC Act are (i) a substantial injury, (ii) that is not reasonably avoidable by consumers, and (iii) that is not outweighed by the benefits to consumers or to competition. Cited examples of unfair practices in the big data context include (1) the failure to reasonably secure consumers’ data and (2) the sale of data to entities that a company has reason to know will use the data for fraudulent purposes.  For instance, in 2006, ChoicePoint, a consumer data broker, settled FTC charges that the company compromised the financial records of more than 163,000 consumers when it furnished those records to identity thieves posing as legitimate subscribers and failed to maintain reasonable procedures to screen prospective subscribers.¹⁰

Considerations for Legal Compliance

The report offers the following considerations for legal compliance relative to the FTC Act: 

• Honor promises made to consumers and provide consumers material information about data practices.
• Maintain reasonable security over consumer data.
• Undertake reasonable measures to know the purposes for which customers are using the company’s data.

Policy Considerations for Companies Using Big Data Analytics

Recognizing the potential for big data benefits and the need to limit possible harms, the Commission report suggests that companies already using or considering engaging in big data analytics should consider:

• Representation.  Consider whether data sets are missing information from certain populations and take appropriate steps to overcome this problem.
• Biases.  Review data sets and algorithms to ensure that hidden biases are not having an unintended impact on certain populations.
• The Accuracy of Predications.  A big data finding of correlation does not necessarily mean that the correlation is meaningful.  Balance the risks of using those results, especially where policies could negatively affect certain populations.  Consider human oversight of data and algorithms when big data tools are used to make important decisions, such as those implicating health, credit, and employment.
• Ethical or Fairness Concerns.  Consider whether fairness and ethical considerations advise against using big data in certain circumstances.  Consider further whether you can use big data in ways that advance opportunities for previously underrepresented populations.

Conclusion

The report concludes with the FTC encouraging multi-stakeholder collaborative efforts to maximize the benefits of big data. The Commission states its intention to continue to bring enforcement actions were appropriate and to highlight and examine big data practices that impact, both constructively and adversely, underserved populations.

Big data is certainly here to stay, and so are the laws that big data practices can potentially violate. Companies in the big data space should review the considerations for legal compliance identified in the report to ensure that their practices do not violate existing laws and expose them to potential liability.

[1] Big Data: A Tool for Inclusion or Exclusion? Understanding the Issues, FTC Report (January 2016), at 12 (hereinafter, “Report”). 

[2] Report at 1.

[3] 15 U.S.C. § 1681a(d). 

[4] 15 U.S.C. § 1681a(f). 

[5] See, e.g., United States v. Spokeo, Inc., No. 2-12-cv-05001-MMM-SH (C.D. Cal. June 12, 2012), https://www.ftc.gov/sites/default/files/documents/cases/2012/06/120612spokeoorder.pdf.

[6] Complaint, United States v. Time Warner Cable, Inc., No. 13-cv-8998 (S.D.N.Y. filed Dec. 19, 2013), https://www.ftc.gov/sites/default/files/documents/cases/131219timewarnercmpt.pdf.

[7] The Commission expressly referenced a statement from its report, 40 Years of Experience with the Fair Credit Reporting Act, and noted that it “does not believe that this statement is accurate.”  Report at 16 n.85. 

[8] Complaint, FTC v. Golden Empire Mortgage, Inc., No. 09-03227 CAS(SHx) (C.D. Cal. filed May 7, 2009), https://www.ftc.gov/sites/default/files/documents/cases/2009/05/090511gemcmpt.pdf.

[9] Complaint, CompuCredit, No. 1:08-cv-1976-BBM-RGV (N.D. Ga. filed June 10, 2008), https://www.ftc.gov/sites/default/files/documents/cases/2008/06/080610compucreditcmplt.pdf.

[10] Judgement and Order, United States v. ChoicePoint, Inc., No. 1:06-cv-0198-JTC (N.D. Ga. Feb. 15, 2006), https://www.ftc.gov/sites/default/files/documents/cases/2006/01/stipfinaljudgement.pdf.

[View source.]