Trying to rent an apartment in New York, San Francisco, or any other high-cost city where housing is in high demand can be very tricky. It doesn't make it any easier when rental listing sites fail to weed out scam artists and tricksters.
Yet when researchers from New York University's Tandon School of Engineering studied more than two million New York City apartment rental listings on Craiglist, they found that the site caught only 47 percent of those deemed to be fake. Suspicious postings lingered for as long as 20 hours before being removed.
While online rental scams aren't new, the NYU study is said to be the first systematic, empirical look at the problem. The NYU team analyzed more than two million rental listings on Craigslist over a five-month period to gain an end-to-end understanding of how such scams are structured and which strategies may undermine them.
"This is the first study to really unpack these rental scams and uncover their vulnerabilities," said Damon McCoy, an assistant professor of computer science and engineering. He said most schemes that utilize credit cards are based in the United States and should be vulnerable to disciplinary or regulatory action.
McCoy also said that the common characteristics of scam ads should make them easier to detect.
"We've shown that rental scams are often built on the same foundation -- there are common templates, emails, IP addresses and other red flags that can be used to develop more sensitive detection techniques in the future," he said.
Common scams include the credit report scam, cloned listings, and phony fees.
In the credit report scam, the would-be renter is told to click a link and purchase a credit report for review by the landlord. The scammer gets a referral commission from the credit reporting site even though there is no property for rent.
In the clone listing scam, legitimate rental listings are copied and posted on Craigslist at a lower price. The scammer asks for a rent deposit via a wire transfer, then disappears with the money.
Phony fees include instances in which victims are asked to pay both an upfront fee and a monthly membership fee to access listings of pre-foreclosure rentals or rent-to-own properties. In the majority of cases, the companies leading the scams have no connection to the properties listed.
McCoy and his associates detected and analyzed about 29,000 fraudulent listings in 20 major cities, ultimately mapping the listings into seven distinct scam categories, most of which involved credit card payments.
They began by developing semi-automated detection techniques that proved highly effective at identifying potentially fraudulent listings. They isolated suspected scams based on shared characteristics; for example, common email addresses, postings with email addresses previously reported in connection to scams, or listings that appeared on other rental sites with different pricing or contact information.
Using an automated "conversation engine," the team engaged some suspicious posters by email, which yielded another set of common features associated with scam communications, namely a set of keywords and personal circumstances -- the responders always claimed to be out of the country -- as well as common IP addresses and embedded links that prospective renters were asked to click.
McCoy, along with Elaine Shi, an assistant professor of computer science at Cornell University, and Youngsam Park, a doctoral student at the University of Maryland, presented the findings of their paper, "Understanding Craigslist Rental Scams" at the Proceedings of Financial Cryptography and Data Security Conference in Barbados this month.