News

Picture this: You’re the landlord of a multifamily building with rents that are affordable for the neighborhood. Most of the people vying for your units submit applications that include possible red flags: a criminal record, low credit scores, or even a past eviction. How do you decide who to rent to?

In our current housing crisis, understanding how landlords interpret these records to choose tenants can provide valuable insight into who can ultimately access shelter and who cannot, especially among renters with limited options. New research by Duke Law Professor Sara Sternberg Greene, co-authored with Barbara Kiviat of Stanford University and Hesu Yoon of the Polytechnic Institute of Paris, sheds new light on this decision-making process. They find that both large and small landlords draw on cultural scripts about responsibility and morality to evaluate prospective tenants, and interpret the myriad forms of data available to them (including what are known as “negative records,” such as a prior bankruptcy or a criminal record) using those beliefs and their own understanding of the law.

“Landlords operate with broadly shared cultural understandings about the nature of risk and the morality of various types of negative records, but with different conceptions of what it means to make rental decisions legally and fairly,” they write in “Getting to Home: Understanding the Collateral Consequences of Negative Records in the Rental Housing Market.” 

Greene, the Katharine T. Bartlett Distinguished Professor of Law, and her co-authors conducted 88 interviews with landlords, property managers, rental company executives, and executives at tenant-screening companies. They focused on two field sites: San Jose, California, a state in which several cities have passed Ban-the-Box ordinances prohibiting landlords from considering criminal records when evaluating tenants, and Durham, North Carolina, where such bans are not part of the state’s legal landscape. Through searching different websites, they created a sample that represented all segments of the rental market, and landlords of various sizes: Some were small individual owner-operators, others were at large companies with multiple properties (often across different states). A little more than one-fifth of their sample were executives. 

Greene knew that landlords today, regardless of their size, have an abundance of data at their fingertips. All of the people they spoke to, she found, sought to be as fair as possible in evaluating it. Some of the tools they use, such as credit scores, are numerical ratings that represent someone’s history of debt and repayment. Others, such as algorithms, ostensibly offer a similarly objective analysis of a tenant based on a set of criteria that executives deem relevant. 

In their interviews, they saw that even those theoretically objective scores and evaluation processes are subject to human interpretation, and even bias. For example, many landlords, both large and small, described having overlooked medical or student debt in someone’s credit history because they saw such debt as a product of bad luck or the admirable instinct to improve one’s opportunities through investing in education. By contrast, they were suspicious of people who had accumulated or defaulted on credit card debt, interpreting it as a sign of irresponsibility. 

“A lot of smaller and larger landlords said, ‘I don't care about medical debt; you have to get yourself help, and we all know hospitals are notorious for overcharging. But credit card debt, you know, that's serious,’” Greene recounted. “But a lot of times when people have a medical emergency, they have to use their credit cards to cover expenses, and if they're not working for a period of time, they can fall behind.” 

They found a similar approach when it came to the people designing the algorithms that the firm used to analyze tenant’s applications. Employees at tenant screening companies, who sell this software to large landlords, told Greene and her colleagues that they’d had clients choose to exclude certain kinds of debt, even though data indicated that including it would make their screening process more accurate at predicting which tenants would not pay rent on time. 

“Some tenant screening companies told us, ‘We even told landlords, Hey, I have some data that shows that medical debt does have some predictive value,’” Greene recalled. “And the landlords of these large companies just said, ‘Thanks. Actually, still don't count that.’”

Cultural understandings color interpretation of data

The findings from these interviews underscore the subjective nature of supposedly objective tools like algorithms. “Algorithms used in the rental housing context are really shaped by cultural understandings,” Greene said. “We think about them as being neutral, and that they couldn’t lead to any kind of bias, but they’re culturally overlaid with all these ideas that we have about good debt and bad debt and fault and no fault, and these American ideals of pulling yourself up from your bootstraps.”

Cultural understandings again come into play when landlords apply their interpretations of the law and fairness as they evaluate tenants. 

Small landlords showed flexibility when a tenant was able to explain a negative record in what Greene and her colleagues call “culturally salient terms,” by telling a sympathetic story and making a connection with the landlord. Employees at large landlords, who were often subject to more formal training on legal compliance and had greater awareness of Fair Housing laws, didn’t feel they could offer the same flexibility. 

One gave Greene and her co-authors an example of having to turn away a prospective tenant who had debt because they’d recently left an abusive relationship and were responsible for two kids, explaining, “You know, as much as I feel for you and your circumstance, because of Fair Housing laws, I have to screen you the same as everyone else.” But large landlords did have their own sets of rules which allowed them to take on tenants with certain negative records if the tenant could provide two months’ security deposit, compared with the normal requirement of one month’s. 

Another area of variation was in the use of criminal records. Many landlords, small and large, in California said they did not consider criminal records because doing so might be legally questionable, thanks to Ban-the-Box laws (a statewide law, the Fair Chance Act, forbids employers to consider criminal records; several cities, although not San Jose, have similar ordinances for landlords).

But others also seemed to have taken the inherent forgiveness in such laws to heart, saying in interviews that a crime in the past shouldn’t be held against someone in the present as long as they’ve served their time. Landlords who did check criminal records were often interested in the type of crime and how recent it was; several cited violent crimes, crimes of a sexual nature, and recency of the crime as relevant for the safety of other tenants.

These findings have important implications for housing policy and for institutions that help tenants. Housing nonprofits, for example, can use these insights to counsel people on how to explain their records in sympathetic and culturally salient ways, and encourage them to reach out proactively to smaller landlords to provide this context. 

Ban-the-Box laws in California may not only make landlords hesitate to use criminal records out of regard for legal compliance, but it also seems to have encouraged people to think about why some localities have banned these records to begin with. “People start to think, 'If it’s the law that this is not allowed, there's probably a reason,'” Greene said. “You could imagine a world in which similar laws prevent consideration  about other types of debt; it could change the culture.” 

Their findings also suggest that policies such as California’s Assembly Bill 12, a 2024 law aimed at making housing more affordable by capping security deposits at one month’s rent, might inadvertently make it harder for people with negative records to overcome a landlord’s hesitation by coming up with additional financial capital for a sizable deposit. 

But Greene notes that relying on the private market to provide adequate affordable housing, especially to people with low incomes and debt on their records, has its limits. “This study really does call into question what obligations our government has to provide housing for people who fall to the bottom of landlords’ preferences,” she said, observing that the cost of housing in some parts of the country has skyrocketed in recent years. 

“What are our obligations as a country to people who owe a lot of debt, often because they experienced a financial shock, or can’t make ends meet due to wages that haven’t kept pace with the cost of living? Who can't pay their utilities because it was freezing and their heat bill was $400 a month, and paying it would take away food for their children? How much can the private market do for them?”