This story was supported with funds from Spotlight DC—Capital City Fund for Investigative Journalism.

Some nights, Katina Butler would listen to her daughter sleep.

“It sounds like she has sleep apnea,” she says.

Butler suspects her 3-year-old’s nighttime wheezing is due to the mold she’s seen in her five-bedroom Southeast apartment. That, coupled with the violence that happens outside her door, prompted Butler to search for a new home for her family of six. She couldn’t believe her luck this past summer when she found the perfect seven-bedroom home on Alabama Avenue SE.

“It’s close to my mom’s place and it’s across the street from the church that I got married in, so it feels like it was meant for me,” she says.

But Butler waited for six months to move into the new house, leaving her feeling frustrated after months of delays at the D.C. Housing Authority, which issues the housing voucher Butler uses to pay her rent.

“I’m just praying that I don’t spend another Christmas here,” she told me in November.

But while Butler was stuck in a moldy apartment, DCHA approved multiple increases in the amount of rent the agency pays her landlord, Roshonda Josephs. (Housing voucher recipients pay a fixed amount of their monthly income in rent, and the government covers the rest.)

I had initially called Butler to inquire not about the conditions in her apartment, but to ask about an online ad for a five-bedroom townhouse at her address advertised under the name “Sharon Rogers.” The advertised rent was $9,322.

“Not only is that not the correct rent, the building is not even a townhouse, it’s an apartment building with three units,” Butler said, chuckling. “No one is paying $10,000 to duck and dive bullets in this neighborhood.”

Butler says she doesn’t know who Sharon Rogers is, but she says that Josephs has received substantial increases in the voucher payments for her unit from DCHA since the agency implemented its “rent reasonableness” tool last July. Butler shared her payment records along with rent determination notices from DCHA, which show that her voucher payment increased from $2,653 to $4,500 in October. Just one month later, Josephs received another increase, raising the monthly rent to $5,344.

“How in the world does she keep getting increases?” Butler asked when we spoke. She asked the same question of Josephs after she noticed the increasing dollar amounts on the management company’s tenant portal. Josephs never responded, and Butler says her access to that tenant portal was later restricted. Josephs, who is the sole beneficiary of the trust that owns Butler’s building, did not respond to multiple requests for comment.

The alarming answer to Butler’s question involves DCHA’s new rent reasonableness tool that uses an algorithm to determine how much a given rental unit is worth.

The process goes like this: Landlords propose rental rates to DCHA, and the agency is then required to conduct a “rent reasonableness” analysis to ensure they aren’t overpaying. The analysis compares the landlord’s requested rent with rates for similar nearby units, called comparables.

But a concerning trend emerged following the tool’s implementation last summer. Throughout hundreds of listings for rental units on affordablehousing.com, Bright MLS, Zillow, and hotpads.com, landlords are advertising units at inflated prices, and in some cases are advertising units that do not even exist. Those ads are sometimes being used as comparables by DCHA’s tool, resulting in potentially unwarranted rent increases for some landlords.

The U.S. Department of Housing and Urban Development is in the midst of an investigation into whether DCHA is overpaying landlords in its voucher program, according to the Washington Post, which previously reported that the housing authority had been doing exactly that.

DCHA contracted with Nan McKay & Associates, which is an “authorized re-seller” of affordablehousing.com’s rent reasonableness tool that the agency uses to determine rental rates for voucher recipients, according to a housing authority spokesperson. The $101,475 contract was inked in May 2023, two months before Nan McKay consultant Dorian Jenkins was appointed as DCHA’s interim director. (Keith Pettigrew was named permanent director in September.)

Asked to describe how DCHA independently verifies the accuracy of affordablehousing.com’s valuation, agency spokesperson Alison Burdo instead explained via email how the tool draws from a database of more than 150,000 units in D.C. and “considers multiple characteristics to identify private market rental property units within a specific geographic radius that are similar to subject units.” 

Property owners can either accept, reject, or appeal affordablehousing.com’s valuation, Burdo says. Once the landlord accepts a rate, DCHA inspects the unit to confirm its condition. But seven out of eight certifications provided to City Paper via the Freedom of Information Act were for rental units whose conditions were listed as “unknown.” 

Burdo says that if a listing is flagged for low reliability or validity, affordablehousing.com—not DCHA—manually reviews them. And while she says that monthly “spot checks on rent reasonableness reports” ensure the accuracy of the algorithm’s recommendations, it’s unclear who conducts such checks, DCHA or affordablehousing.com.

Richard Cupelli, the founder of affordablehousing.com, says that the rollout of the new rent reasonableness tool in D.C. has been challenging. Cupelli says that affordablehousing.com “is a source, but it’s not a primary source” of comparable rental units. The company could not produce statistics on the number of analyses that used data from leases or rent rolls.

Landlords can generally list units online at whatever rate they want, and when the algorithm treats that inaccurate data as a primary and credible gauge of the rental market, it can lead DCHA to approve exorbitant rates for units rented to voucher holders.

DCHA’s lack of transparency and oversight with its rent reasonableness process has resulted in the approval of inflated rates as well as confusion for landlords and tenants alike. While some landlords are being overpaid, voucher holders are deceived by ads for apartments that don’t exist.

“The owners are used to getting above market rents and they have been trying aggressively to add data,” Cupelli says, adding that “the owners in D.C., they’re definitely trying to manipulate the system, so it’s been a battle.”

***

London’s building at 4520 Georgia Ave. NW, for example, contains only 12 units, according to its certificate of occupancy, but there were 24 listings on Bright MLS, including 12 listings for four-bedroom units. On affordablehousing.com, London advertised 38 units for the building. Affordablehousing.com later distributed these ads to other websites like Zillow and hotpads.com as a benefit of a premium membership, which London purchased.

According to the D.C. Department of Housing and Community Development’s database of rental units, called FileNet, the most recent registered rents in the Georgia Avenue building in 2023—eight one-bedrooms and four two-bedrooms—range from $943 to $1,464. (All rental units in D.C. are required to be registered with DHCD.)

But on affordablehousing.com, the advertised rents ranged from $2,735 for one-bedroom units up to $6,281 for nonexistent six-bedrooms. On Bright MLS, London advertised four-bedroom units, which do not exist in the building, for $6,620 to $6,995.

The same discrepancies existed for some of his other buildings, including 24 ads for an eight-unit building at 5348 Call Pl. SE; eight ads for a four-unit building at 2887 Hartford St. SE; 15 ads for the 14-unit building at 2435 Ainger Pl. SE, and 15 ads for a six-unit building at 325 34th St. SE. (Most of the listings disappeared after City Paper’s interview with London.)

In a phone call, London explains that he regularly posts four listings per unit for marketing and search optimization purposes and never intended for the duplicates to be used as comparables.

But in response to questions about London’s multiple listings, Christine Reap, director of corporate communications at Bright MLS, writes in an email that, “when Bright MLS’ subscribers (all of which are licensed real estate professionals) submit property information to us, they must submit accurate information. We invest significantly in producing the most accurate, reliable, up-to-date real estate data anywhere. And our data compliance team investigates all reports of inaccuracies and is taking action to investigate this report to ensure that our system is accurate.”

London adds in an email that he has heard property owners refer to advertised rents “on their voucher listings for older rent-control buildings as ‘placeholder rents.’” 

“The rent is not a meaningful number in that the actual rent is determined in the future when the lease-up is processed but the system still requires a rent to be entered,” he adds. “If the comps are filtered properly, these placeholder rents will not be part of the rent determination.”

Burdo says in most cases, the rent reasonableness review has led to a decrease in approved rates. And Michael Lazdowsky, managing director of affordablehousing.com, says via email that their internal analysis of 3,600 valuations from July 2023 through Jan. 30, 2024, show an average approval of $263 less than the monthly rates landlords initially requested. But, Burdo acknowledges, London’s fake ads for 4520 Georgia Ave. NW may have been used as comparables if they were recently listed online.

Ashley DeRosa, a property manager with Birch Management, says that while approved rents may have gone down in Northwest, “the rents have increased exponentially in Southeast.” 

In August, 75 percent of ads for apartments with four or more bedrooms on affordablehousing.com belonged to London. So when Josephs, a former vice secretary of the DC Housing Providers Association, requested an increase in the rent DCHA paid for Butler’s apartment, London’s online ads were the only source of comparable data used by the algorithm.

DCHA’s September 2023 rent assessment for Butler’s apartment selected three five-bedroom apartments from London’s building at 325 34th St. SE as comparables. The advertised rents ranged from $5,337 to $5,352. (The three comparable units were selected from a pool of nine in London’s building that only contains six total units, according to the building’s registration with the city.) The algorithm determined that a 50 percent increase—from $2,653 to $5,344—was reasonable according to the rent determination notice shared by Butler. DCHA approved the payment.

Quanii Garrish, a former tenant at 325 34th St. SE, was incredulous when I told her that a five-bedroom unit in the building was advertised for more than $5,300. She paid $2,400 per month with a voucher for a five-bedroom unit until she moved out in 2021. She describes London as a caring landlord who provided gift cards to tenants at Thanksgiving, but the increase in rent shocked her.

“It will never be worth $5,300 to live in that neighborhood—ever,” says Garrish, who moved out after a stray bullet came through her window and the police broke down her door.

***

“When an owner is trying to take advantage of the system by adding a bunch of listings on the front end, it’s obvious that that’s occurring,” Cupelli says. “We have processes in place to see who they are and they could potentially be banned.”

London says in an interview that during a landlord training session in June of 2023, an affordablehousing.com employee advised landlords to list all of their existing units—vacant and occupied—on the website in order to populate the database. Then in December 2023, affordablehousing.com corrected their guidance with instructions to list all units according to their availability, London says.

A comparison of rent reasonableness valuations and executed housing voucher contracts, obtained from DCHA via FOIA, shows that only London’s internet listings were used as comparables for some of his own units that are rented to voucher holders, leading DCHA to approve significant rent increases.

London requested $6,705 in rent for a five-bedroom unit at 2435 Ainger Pl. SE, which previously rented for $3,280, according to London’s filing with DHCD. The algorithm selected three units advertised by London in the same building as comparables. The data sources for two of the comparables were cited as “internet listings” while the third is labeled “owner-added.”

The rents for the comparable units were listed at $5,775, $5,780, and $5,800, which line up with the rents advertised on BrightMLS last summer. The algorithm ultimately found that the five-bedroom unit in question was worth $6,296, in part because the utilities would be paid by the property owner. But that figure is about $400 above DCHA’s “payment standard” of $5,897. (The payment standard is the maximum amount that DCHA will pay for a given unit.)

London denies receiving a rent increase for the unit, but a notice of recertification he submitted in his October 2023 registration with the DHCD, lists the amount ultimately approved by DCHA: $5,897. A redacted voucher contract, provided by DCHA via FOIA, lists the same approved amount.

The current tenant, who moved into the building in August of 2023, confirms that their voucher covered $5,897 in rent. The tenant, who describes London as “the best landlord I’ve ever had,” says they are paying London $100 a month to pay off a security deposit tied directly to the rental rate that DCHA approved. When asked about the security deposit, London denies he’s collecting the monthly payment. “The tenant is not out one penny,” he says. 

Increases for London’s other voucher payments that were approved after the tool came into use can be traced to the same pattern of inflated rents from internet listings.

A valuation for a four-bedroom unit at the building on Ainger Place SE used the rents from three internet listings from the same building as comparables, each advertised for $4,432. The tool determined that the reasonable rent for the four-bedroom unit was $4,663 because, unlike the comparable units, the owner would be paying some utilities. DCHA approved the recommended rate according to a voucher contract provided by DCHA via FOIA. The previously registered rent was $2,602.

A valuation for another four-bedroom unit in the building shows the same results using three units advertised online for $4,277. The algorithm recommended a rate of $4,614, which DCHA ultimately approved. The previously registered rent for that unit was $2,797.

The tool’s valuation of a three-bedroom unit in London’s building at 2908 30th St. SE used three internet listings from the same building, each advertised for $3,484, as comparables. A voucher contract provided by DCHA and a rent determination notice included in London’s DHCD filing for the building, shows an approved rent of $3,494—an increase of $1,234 from the previously registered rent of $2,260.

A comparison of London’s DCHA-approved rent increases and the rental rates he registered with DHCD in June of 2023 shows that rates for only a small handful of his units stayed the same or slightly decreased after the rent reasonableness analysis. For the most part, the rent payments approved by DCHA after the implementation of the algorithm increased by more than 50 percent. Of the 14 units reviewed, 10 were approved for increases after July of last year. At the Ainger Street building, approved rents increased by an average of 62 percent. At 2887 Hartford St. SE, the approved rent for a three-bedroom increased from $2,129 to $3,655, according to London’s DHCD filings. Some of London’s units at 3070–3074 30th St. SE had approved increases of more than 70 percent, with one six-bedroom apartment going from $2,797 to $5,107—an 82 percent increase.

London, for his part, denies receiving any rent increases. After our interview, he says he contacted DCHA to remove his rent-controlled units from the comparable database. He adds that affordablehousing.com “told us they had millions of data points and they had entire companies sending them rent rolls. They made a point of emphasizing that they do not use marketing in determining rent reasonableness, and that they use actual non-subsidized market rate … in-place leases.”

***

But Brandon Lwowski, senior director of research at HouseCanary, a national technology and data-driven real estate brokerage firm, says that while there is a benefit to leveraging online rental ads, the data can be unreliable.

“If they’re just scraping from open marketplace listings, there can be some price inflation in those numbers,” he says. “The landlord may inflate those numbers to gamify the algorithm.”

He cautions: “When you depend on that list price as your honest truth, your model will become biased and will inflate rental prices as well because it’s learning from these listings.”

When asked, Cupelli sounded surprised by his algorithm’s use of owner-submitted internet listings as comparables because “they have very little weight.” He speculates that “it’s possible that if the owners provide information to the agency, the agency could choose to use those,” but “if this type of activity is occurring, we’ll need to monitor it more.”

But in a 2020 podcast interview with real estate investor Jason Hartman, Cupelli describes one of the benefits to landlords of listing an apartment on affordablehousing.com (formerly known as gosection8.com) and explains how the company’s online advertising and rent reasonableness tool work together.

“You have to get an inspection [to] make sure that the rent that you’re asking is reasonable,” he said on the podcast. “We built a method to streamline that. So the rent reasonableness functionality in gosection8, you’re much more likely to get your rent approved, if you use gosection8 because there is the ability for you to add comparable data and defend your rent through the back end system that we created, which is used by the agencies.”

The algorithm’s susceptibility to potentially faulty online ads is not its only flaw. HUD guidelines require that only market units be used as comparables (as opposed to rent-controlled units or those receiving government subsidies). Because all of London’s units are rent-controlled, their use as comparables violate HUD’s directive.

A DCHA spokesperson says that the rent reasonableness algorithm attempts to exclude rent-controlled units by identifying “listings with statistically significant low rents and prevents them from being included as comparables.” But because some landlords like London advertise rent-controlled units for prices far above the rents registered with DHCD, the algorithm could fail to exclude them.

Cupelli says that identifying rent-controlled units is difficult. But in addition to looking for lower rental rates, the tool can also flag buildings that were built before 1975 or if they contain five or more units—two attributes that indicate rental units could be rent controlled. Affordablehousing.com also maintains their own list of rent-controlled units, he says.

According to Burdo, affordablehousing.com provides instructions and training to landlords that subsidized units should not be submitted as comparables. In addition, DCHA provides addresses of subsidized units to affordablehousing.com so they are excluded as comparables.

But while some landlords are benefiting from the tool’s flaws, many others are confused by the lack of transparency.

“No one knows what they’re doing, not even DCHA,” says landlord Peter He.

The rent reasonableness analysis for his market-rate, six-bedroom row house at 603 21st St. NE used rent-controlled units from London’s building at 325 34th St. SE as comparables. London’s two six-bedroom units were listed online for $5,652 each. The third comparable unit used in the analysis was the exact same six-bedroom house for which He was seeking a rent reasonableness review.

According to He, the analysis led to an approved rent of $5,835, a $400 decrease for his rental unit, and he’s struggling to understand why. He asked a DCHA employee for the rent reasonableness valuation for his unit after I contacted him. But He says the employee did not know what a rent reasonableness valuation was.

Finding a market-rate comparable unit is almost impossible, He tells me. “In certain zip codes, there is no market base,” he says. “It’s all vouchers.” He is perplexed by how or why London’s rent-controlled apartments in Southeast were used as comparables for his market-rate row house in Northeast.

Even as inaccurate online ads create problems for affordablehousing.com’s tool, the website actually contributes to the proliferation of those ads. The site’s “listing multiplier” promises landlords who purchase a premium membership that “each rental listing you create on affordablehousing.com is automatically distributed to other real estate sites to maximize your exposure and drive more leads.” 

London, who purchased a premium membership, denies listing ads on hotpads.com when I ask about the 32 ads for 4520 Georgia Ave. NW found on the website. He explains that affordablehousing.com syncs the data from its own website with other online rental websites.

Lazdowsky confirms in an email that the company lists ads on other websites as a paid service for landlords.

DCHA is working to remedy the issues raised in this article and is recalculating London’s rent reasonableness valuations, according to Burdo. But London’s listings have potentially affected other landlords’ rent reasonableness determinations, and Burdo says the agency will recalculate rent reasonableness determinations that used comparable units listed by London if necessary, but she could not provide a number.

“DCHA is utilizing the quality control checks within the AffordableHousing.com software and is collaborating with the company to develop additional tools tailored to DCHA’s unique needs to ensure questionable listings are tagged for AffordableHousing.com to review,” Burdo writes in an email. “DCHA is implementing quality control processes, including the manual review of a random sample of rent reasonableness valuations. That sample accounts for roughly 5% of valuations,” she adds.

Lazdowsky says in an email that affordablehousing.com is “initiating steps to enhance our collaboration with DCHA to help prevent the manipulation of rent stabilized units in our comparable database,” including identifying landlords who are manipulating market rents through false listings and excluding those owners’ data from being used as comparables. Owners will also be required to indicate if their property is rent controlled going forward, he says.

Lazdowsky adds that they will be “reaching out to DHCD to explore a data sharing agreement, aiming to more accurately identify rent stabilized units in Washington, DC.”

It remains to be seen how these changes will pan out and how they will be communicated to landlords and tenants.

As for Butler, she received the news that her new house passed inspection a few days before Christmas, and her family finally moved in on Dec. 31. Her new landlord allowed her to move in without paying a security deposit as she awaits assistance with that payment from the D.C. Department of Human Services. London’s tenant at 2435 Ainger Pl. SE, however, says they are still slowly paying off a security deposit of $5,897.

Correction: This story incorrectly reported that DCHA negotiates rental rates with property owners. It has been updated to say that landlords have the option to accept, reject, or appeal the agency’s proposed rate. The story also incorrectly attributed information to an agency spokesperson. This version has been corrected.

Featured Local Savings