We have become inured to a world of surveillance so pervasive it would make the Stasi blush. Much of this infrastructure is built on our nominal consent in the guise of consumption choices. We carry around tracking and recording devices in the form of “phones” because they also contain navigation tools, music libraries, messages with our intimates, games, cameras, and a huge variety of other tools to make our lives more convenient and connected. We accept that our online lives will be monitored, not always thinking of it, because doing so makes it possible to provide many services for free and makes it easier to find things and people that fit one’s idiosyncrasies. And, as brick-and-mortar stores close and more people stay in touch with each other through networked communication devices, it is increasingly difficult to live one’s life without “opting into” a surveillance architecture. Many (most?) of us would rather that the conveniences and connectivities of modern life not be connected to a network of surveillance—especially as the second Trump administration knits together these networks of commercial surveillance even more closely with state surveillance and repression–but we find ourselves feeling powerless to do much about it.
We association most of the modern infrastructure of nominally opt-in surveillance is associated with the rise of Big Tech, but as Nakita Cuttino’s new article The Presumption of Creditworthiness reminds us, before Big Tech came credit reporting. Over the second half of the Twentieth Century, credit reporting agencies developed the basic approach of collecting data that businesses had on their customers without customer consent and compiling into files that other businesses and law enforcement agencies could buy. Once Fair Isaac Corp. developed its initial credit scoring model, the credit reporting industry also became the first to sell its data to firms with proprietary models that could be used to automate customer evaluation and, eventually, to segment consumer markets (and to target vulnerable customers with the most predatory deals). And as consumer credit became a core part of American life, the data collected by these companies became increasingly valuable for all kinds of businesses (employers, landlords, insurance companies) and the difficulty of opting out of the surveillance dragnet became increasingly high.
Cuttino’s thorough and thoughtful article gives a sense of how difficult opting out of the surveillance dragnet can be. Cuttino’s subject is “unscored” or “credit invisible” consumers—those about whom credit reporting agencies have insufficient information to produce a credit score. Approximately 32 million US adults—that is, nearly 13% of the adult population—are in that category. In the current system, unscored individuals are mostly presumed to be of low creditworthiness. Cuttino argues convincingly that this is a problem, and she provides a creative but tractable account of how it could be solved by flipping the presumption. In the process of doing so, she also provides fodder for a broader critique of using credit scores to parcel out opportunities and, indeed, of the creeping inevitability of megacorporations and governments using statistical analysis of our past to carve out the future paths available to us.
Cuttino points to several problems with treating unscored customers as presumptively unworthy of credit (or of only high-cost credit). The first is lack of credit history (or stale credit history) is not associated with nearly as high a risk of nonpayment as the current system assigns to them. Multiple studies indicate that most people with limited credit history are quite reliable payors. If they are being treated as if they are not, creditors are inefficiently restricting credit. Because Black and Latine individuals are disproportionately likely to be unscored—because, to oversimplify, people from these communities are less likely to have relatively wealthy and credit-sophisticated family members to ease them into credit markets—the inaccurate presumption of non-creditworthiness reinforces, as it also reflects, existing racialized inequalities. As Cuttino puts it: “the inadvertent creation of unscored consumers frustrates the credit scoring system’s promise to mitigate discrimination and promote equity” by providing ways of evaluating creditworthiness that rely on predictive data rather than stereotypes. (P. 469.)
In addition, the consequence of treating individuals without credit history as presumptively risky is to create a self-fulfilling prophecy. Unscored individuals find themselves either shut out of mainstream markets or offered credit on terms that are more difficult to pay. Borrowers with riskier loans have higher risks of nonpayment—loan risk becomes borrower risk—and their potential need to borrow again to fill gaps in income exacerbated by high payments. Borrowers deprived of credit might find themselves deprived of opportunities to build wealth or just earn income—by acquiring a house that itself appreciates in value or a car that helps them get to a job—which, among other things, makes it more difficult for them to pay future loans. And because credit scoring is used outside lending markets, unscored individuals will also find other opportunities pulled away from them and/or accessible to them only at higher prices. Since being unscored is largely a result of not being part of a community with sufficient resources to underwrite one’s access to credit markets, attaching these cycles of disadvantage to lack credit history is likely to add further cycles to the reproduction of inequality that credit reporting was meant to help remedy.
After diagnosing the problem, Cuttino then explains why the market solutions currently on offer will not suffice. The two most promising categories of market solutions Cuttino examines (she examines four total) are alternative underwriting and alternative lending.
Alternative underwriting draws on data from the expanded surveillance dragnet of our modern world to attempt to sort out the risky from the less risky unscored borrowers. Alternative models have shown some promise—especially when they make use of consumers’ payments on accounts not tracked by credit reporting agencies—but they are not great predictors and they come with risks. One risk is that requiring unscored borrowers to provide more information about themselves makes them more vulnerable in the case of a data breach and also requires borrowers to submit themselves to more pervasive surveillance and judgment regarding often intimate details about their lives. This risk further entrenches the norm of mass surveillance and makes it even more difficult to opt out of. Another is that some of the sources of information drawn on reinforce existing inequalities—borrowers that attended a historically Black college, for example, have been treated as riskier in at least some models.
The other promising market solution, alternative lending, involves products like “earned wage access” and “buy-now, pay-later” options. By these means, lenders provide credit to borrowers after at most a limited credit check, often because they require borrowers to link to payroll and/or bank accounts. While these products do provide sources of credit to at least some unscored customers at relatively low cost, they usually do not help borrowers build credit history (since they do not usually report to credit reporting agencies). And they also require customers to disclose more sensitive information than most customers. Most worryingly, they too-often serve as gateway drugs to higher cost credit products.
What makes the problem of unscored borrowers so vexing, Cuttino argues, is that treating lack of credit history as equivalent to a history of difficulties with repayment produces a form of adverse selection. “Only the most ‘privileged’ of quality unscored consumers will opt out, such as those who can operate solely in cash or who borrow from relatives. The remaining unscored consumers will be compelled to opt in,” which will then force them to use credit products that are more difficult to repay, that require them to expose more of their lives to lenders, and/or that do not actually help them build credit history. (P. 492.)
Adverse selection problems are, of course, problems of inadequate information. But because Cuttino is attuned to the costs that come with acquiring information about individuals’ lives, she is attuned to the limits of market solutions that would solve the problem by increasing surveillance. A more satisfactory solution, she argues, would be using one or more mechanisms to mandate a presumption of creditworthiness. “Such a policy might mandate that consumers who are quality credit risks subsidize those that are poor credit risks to ensure a market exists for those who are unknown risks—i.e., unscored consumers. Alternatively, it might mandate that lenders extend introductory mainstream credit on favorable terms to unscored consumers while absorbing—or being subsidized for—the associated risk.” (P. 493.)
The last section of Cuttino’s article provides some visionary ideas for making this policy goal real, combining interpretations of existing law (primarily under the Equal Credit Opportunity Act) and supplementing with new statutes. Interested readers should turn to the article for the details.
To me, the most interesting and generative aspect of Cuttino’s account is how it provides a wedge into a more deeply critical analysis of the logic of credit reporting. Throughout her article, she points out how increasing the information available to lenders does not necessarily increase predictive accuracy and, even when it does, it does not necessarily reduce the racially disparate impact of credit models. Moreover, more than most scholarly observers of consumer credit markets in the modern world, Cuttino is constantly alive to the risks and costs that come with providing more information about oneself. These and other observations lead her to advocate for a credit market that uses cross-subsidy (and/or public options) to make lenders less reliant on huge amounts of information to narrowly target their offerings. Cuttino’s focus is on unscored borrowers, but as she acknowledges at several points, the basic insights that she develops cannot be easily cabined to that class of consumers. Future work on consumer credit—and, indeed, on surveillance-facilitated commerce more generally—would do well to develop these insights.
