The Residuals Problem: How Algorithmic Underwriting Creates a Fair Lending Blind Spot for Entertainment Workers
Shreyash Shrestha | May 7, 2026
Actors, writers, and directors, who are covered by SAG-AFTRA and WGA collective bargaining agreements (CBAs) often receive their income through residuals. These are contractually mandated and federally reportable recurring payments sent out each time an individual’s work is redistributed across platforms, networks, or foreign markets.. For many guild-covered workers, residuals serve as their primary or supplementary source of income, (Hiller, 2025). Yet, when these workers apply for mortgages, loans, and other forms of credit, their credit may not be evaluated properly by current underwriting systems.
Particularly, credit lenders have expanded their reliance on fintech partners for algorithmic underwriting, a process that assesses creditworthiness by pattern-matching applicant data against historical borrower profiles built around traditional salaried employees, (Consumer Financial Protection Bureau, 2023). The issue is that these models are made for predictable W-2 wages and may struggle to recognize residual income. The result is not just an inconvenience for entertainment workers, but a fair lending blindspot and a third-party oversight issue that can lead to serious legal consequences for lenders.
Residuals are Not Gig Income
A common mistake while analyzing this issue is comparing the residual wage structure to that of a gig worker’s. Specifically, the entertainment workers at issue here are not simply independent contractors. Instead, they are often covered by guild CBAs that establish them as employed and may hold benefits such as pension plans, health coverage thresholds, and negotiated compensation structures. Residuals are part of this labor structure, (DLA Piper, 2023). For example, SAG-AFTRA operates a union pension fund intermediary for residuals that receives payments from producers and distributes it to members on a structured schedule. This creates a documented disbursement record for each transaction, or in other words, a paper trail that is more auditable than many income sources that underwriting models already treat as standard.
This feature is an important distinction because it renders that if a lender’s model were to fail in recognizing a residual payment as legitimate income, the problem is not that the income is undocumented. Rather, the problem is with the model itself. In algorithmic underwriting specifically, models operate at scale by using standardized income strategies, meaning that unless directly integrated, they have no mechanism to request or evaluate SAG-AFTRA disbursement records, (Legal Clarity, 2024). This is a design failure and would cause the originating bank significant legal trouble under a fair lending challenge: a regulatory or legal claim that a lender’s practices produced discriminatory outcomes.
The Disparate Impact Problem
The issue here is that a plaintiff could use the disparate impact doctrine, a legal theory, to strengthen their fair lending challenge, and under existing precedent, prevail. Specifically, a disparate impact claim allows a plaintiff to challenge a neutral policy for producing discriminatory effects, even without evidence of intent.
To understand why this would work in the entertainment space, there is a two-step connection. First, there is a representational gap within the entertainment industry as showcased in 2024 when people of color made up 44.3% of the U.S. population but only 25.2% of lead roles in the year’s top theatrical releases, (UCLA Institute for Research on Labor and Employment, 2025). Second, there is an algorithmic issue as underwriting models are designed to assess standardized income patterns, meaning income types that fall outside those patterns are treated as irregular. But because the irregular income type here is distributed along racial lines, the outcome can be seen as discriminatory. The fact that this effect exists could establish a prima facie disparate impact claim, and given that SAG-AFTRA's auditable disbursement records, could allow the claim to proceed past the pleading stage..
To provide context, the doctrine stems from the Supreme Court’s 1971 decision in Griggs v. Duke Power Co. In the case, the Court held that the high school diploma requirement violated Title VII, (a federal law prohibiting discrimination) because it disproportionately screened out black applicants. Here, evidence of intent was not required, instead the effect was enough. This precedent has been extended to lending though the Equal Credit Opportunity Act (ECOA) and the Fair Housing Act (FHA) as both prohibit discriminatory outcomes regardless of intent.
Furthermore, the Consumer Financial Protection Bureau (CFPB), the main federal agency responsible for protecting consumers in financial markets, has already put lenders on notice. In 2022, the CFPB made clear that lenders cannot hide behind model complexity, (Consumer Financial Protection Bureau, 2022). In 2023, the bureau went further by using a profession-based income estimation as its explicit example, stating that telling an applicant they had "insufficient projected income" is not a legally adequate explanation, (Consumer Financial Protection Bureau, 2023). This is significant because the failure can be severe in algorithmic underwriting since the black-box nature of these models makes it impossible to audit which specific income classification rule produced the denial.
Building a Compliance Framework
A key note before moving further is that in algorithmic lending, the originating bank cannot avoid responsibility by blaming the fintech company’s model. Under third party risk guidance frameworks developed by the Federal Deposit Insurance Corporation (FDIC) and Office of the Comptroller of the Currency (OCC), the bank that originated the loan faces full legal risk for using the fintech model, (Federal Deposit Insurance Corporation et al., 2023). Regulators have already acted against banks for this kind of failure. For instance, in 2023, Cross River, the originating bank, was using an underwriting model via a fintech partner, and faced a consent order by the FDIC since the agency found out the bank's compliance systems were not built to detect whether its fintech partners were producing discriminatory outcomes, and they faced legal action for it, (Consumer Finance Monitor, 2023). That same year, the OCC’s consent order against Blue Ridge Bank went even further, citing systemic failures in the bank's compliance management systems around fintech partner underwriting, specifically the bank's inability to oversee whether its partners' models were producing compliant credit outcomes, (Office of the Comptroller of the Currency, 2023). These orders establish that inadequate oversight of fintech partner models carries direct consequence.
For banks operating under the same model, this pattern defines the compliance floor. Particularly, this means two things. First, contractual agreements between banks and fintech partners should require the fintech partner to disclose how its model classifies non-W-2 income. In this case, the model would need to include structured disbursements from union intermediaries like SAG-AFTRA's residuals processing unit. If the model does not have a category for that income type and does default it to irregular or insufficient, it must be ensured that the bank is informed before deployment. The FDIC's 2024 Consumer Compliance Supervisory Highlights specifically criticized banks that lacked full access to all the variables used in their fintech partners' underwriting models, (Federal Deposit Insurance Corporation, 2024).
Second, banks should be running periodic model audits that test for disparate denial rates across protected class categories. In this case, if a model produces statistically higher denial rates for SAG-AFTRA members relative to W-2 employees with equivalent debt-to-income ratios, that pattern is detectable before it becomes an enforcement action, (Consumer Financial Protection Bureau, 2024).
The Cross River and Blue Ridge consent orders act as direct examples demonstrating that banks who take preventative action will be better positioned.
Conclusion
While enforcement targeting entertainment workers has not yet arrived, the underlying conditions are directly in place. Algorithmic lending through bank-fintech partnerships is growing and the structural conditions for income misclassification are present. At the same time, the CFPB has already put lenders on notice, recent consent orders have established that banks cannot outsource their fair lending obligations to fintech partners, and the entertainment industry’s demographic makes it a strong candidate for a fair lending issue. The only remaining piece is for someone to put these facts together in litigation.
Entertainment law has spent decades using guild agreements to anticipate harms before courts recognized them, from residuals themselves in 1960 to AI replica protections in 2023. Fair lending is the next area where that same anticipatory legal framework is needed, and the statutory architecture is already there to support it.
References
Consumer Financial Protection Bureau. (2022, May 26). CFPB circular 2022-03: Adverse action
notification requirements in connection with credit decisions based on complex algorithms. https://www.financialservicesperspectives.com/2022/05/cfpb-provides-new-guidance-on-discrimination-in-algorithmic-credit-decisions/
Consumer Financial Protection Bureau. (2023, September 19). CFPB issues guidance on credit denials by
lenders using artificial intelligence. https://www.consumerfinance.gov/about-us/newsroom/cfpb-issues-guidance-on-credit-denials-by-lenders-using-artificial-intelligence/
Consumer Financial Protection Bureau. (2024, July 2). Fair lending report of the Consumer Financial
Protection Bureau. Federal Register. https://www.federalregister.gov/documents/2024/07/02/2024-14533/fair-lending-report-of-the-consumer-financial-protection-bureau
Consumer Finance Monitor. (2023, May 4). FDIC consent order with Cross River Bank indicates
heightened scrutiny of bank-fintech partnerships. https://www.consumerfinancemonitor.com/2023/05/04/fdic-consent-order-with-cross-river-bank-indicates-heightened-scrutiny-of-bank-fintech-partnerships/
DLA Piper. (2023, December 20). Inside the SAG-AFTRA collective bargaining agreement.
Federal Deposit Insurance Corporation. (2024, March 28). Spring 2024 consumer compliance supervisory
highlights. https://natlawreview.com/article/takeaways-fdics-spring-2024-consumer-compliance-supervisory-highlights
Federal Deposit Insurance Corporation, Board of Governors of the Federal Reserve System, & Office of
the Comptroller of the Currency. (2023). Interagency guidance on third-party relationships: Risk management. https://www.fdic.gov/news/financial-institution-letters/2023/fil23029.html
Griggs v. Duke Power Co., 401 U.S. 424 (1971). https://www.law.cornell.edu/supremecourt/text/401/424
Hiller, B. (2025). Hollywood acting industry statistics. Bernard Hiller Acting.
https://bernardhiller.com/hollywood-acting-industry-statistics/
Legal Clarity. (2024). How participations and residuals work in entertainment.
https://legalclarity.org/how-participations-and-residuals-work-in-entertainment/
Office of the Comptroller of the Currency. (2023). Formal agreement: Blue Ridge Bank, N.A.
https://www.occ.gov/news-issuances/news-releases/2023/nr-occ-2023-3.html
UCLA Institute for Research on Labor and Employment. (2025). Hollywood diversity report.
https://irle.ucla.edu/research/emri/hollywood-diversity-report/
