The Financial Services Authority (OJK) has found at least one instance of a single borrower borrowing from up to 40 P2P lending platforms. (B1 Photo)
AI-Powered Debt Collection Can Help Avoid Another P2P Lending Crisis
BY : GUILLERMO MARTIN
AUGUST 14, 2019
Peer-to-peer, or P2P, lending should have been the long-awaited answer to a complicated financial inclusion problem: how do we help the poor escape the poverty cycle?
Finally, there is a commercially viable way to lend money to the "riskier" segments of the market thanks to high mobile internet penetration and the elimination of expensive middlemen.
Sixty-six percent of Indonesia's population was unbanked in 2018 and cash was king. P2P lending platforms, which typically match hopeful borrowers with private lenders, offered the perfect solution to the problem. Regular Indonesians gained much-needed access to credit, while lenders had an opportunity to gain returns higher than many other investment opportunities at the time.
Then, everything went wrong.
Loan Sharks Hiding Behind P2P Lending Mask
In January, P2P lending was the third most-complained about sector in Indonesia. Stories of criminal harassment can still be found on social media grouped under the hashtags #korbanpinjol or #korbanfintech ("victims of online borrowing" and "victims of fintech" respectively) with tale after sordid tale of victims warning against online borrowing.
Borrowers are crushed by impossible interest rates (up to 2 percent a day) and administration fees that lead to ballooning debts by unscrupulous lenders, no matter if their initial borrowed amounts were small. Desperate borrowers then refinance their loans with other P2P lending companies, again and again. They're now trapped in a vicious cycle.
Then comes collection time. Intimidation, sexual harassment, breaches of data privacy, blackmail and harassment of friends and family are all part of the horrifying norm. One of the patented debt collector tactics involves them creating WhatsApp groups, and adding the borrower’s friends, family and coworkers to shame delinquent borrowers. In these groups, borrowers are branded "fugitives" that must be hunted down. Debt collectors will often demand that members in these groups reveal where the borrowers are "hiding."
Unfortunately, even legitimate companies may fall on desperate times due to high default rates, and without a viable collection strategy planned, may unknowingly employ third-party debt collection agencies that utilize these barbaric tactics to chase their debts.
One of the victims of such harassment was a Jakarta taxi driver who committed suicide last February after failing to pay outstanding loans from 20 different lenders.
His suicide letter contained a plea for the Financial Services Authority (OJK) to eliminate online lending, which he characterized as a "devil’s trap."
The taxi driver's case showcases two of P2P lending's biggest problems – harassment of borrowers, and borrowers inevitably defaulting on their loans if they borrow from too many lenders. The OJK found at least one instance of a single borrower borrowing from up to 40 platforms.
The OJK has attempted to regulate the market, but has come face-to-face with the uncomfortable truth: the internet is impossible to regulate. I can speak to these problems in Indonesia with some first-hand knowledge, but we are hardly unique.
Lessons We Haven't Learned
China's issues with P2P lending more often stemmed from loan defaults that forced even higher interest rates and the closing down of P2P lending platforms, and from taking investors' life savings with it.
The Philippines, another infamous recipient of P2P lending, faced issues that ring closer to Indonesia's. So too, did Vietnam.
The trajectory, though, is always the same.
P2P lending gains significant attention for providing "a real solution," and investors begin pumping funding into these platforms. The industry is becoming marred by bad players ridiculous fees. A combination of that and no real debt collection strategy leads to increasingly desperate lenders. Borrowers begin to report harassment by lending platforms. Lives are lost.
Regulators have scrambled to stop the situation from getting worse. Now we come to an important question: could we have prevented all of this?
Borrowers Need to Learn Financial Basics
Low-income individuals often can't quite grasp the idea of interest rates, making them easy pickings when they are sold on weekly installment schedules. If lenders take advantage of them, they can't pinpoint the wrong done to them, or what they can do about it.
In fact, those earning a lower income may not even be equipped with money management skills necessary to handle debt, which may contribute to higher default rates, and an inability to figure out real solutions to problems caused by their debts other than refinancing from lenders of ill-repute.
Any social good that could have been felt from increasing access to financial products is undercut by the lack of knowledge on how to truly maximize these offerings.
KPMG has noticed the issue as early as 2017, and today this rings truer than ever. Educated borrowers are better equipped to protect themselves against bad lenders, and more importantly, can make decisions that will actually benefit their long-term financial standing.
Credit Checks: Necessary Evil?
Credit checks were the very reason behind P2P lending's necessity, but the industry's failings may sometimes remind us: there was a reason why they were necessary in the first place.
The P2P lending industry needs to conduct robust credit checks, and they need to do so without excluding previously underserved segments from accessing the market.
Fortunately, third-party alternative credit scoring solutions have been launched to bridge this important gap. Solutions like smartphone-based credit scoring solutions utilize robust artificial intelligence to gain information about the creditworthiness of a candidate just through their smartphone and could help P2P lenders provide financing on fair terms to borrowers whom traditional systems have failed.
There are also artificial intelligence-powered solutions to collect debts, too.
Ethical and Personalized Debt Collection
Companies like AsiaCollect strive to help companies maximize their non-performing loans, from offering credit management advisory and Software-as-a-Service (SaaS) solutions, all the way to purchasing debt portfolios.
AI and machine learning can be used to analyze the behavioral and emotional psychology of borrowers, thereby enabling call center operators to communicate more effectively with different personality types. Our platforms are also able to identify the best times and channels (SMSes, emails, social media) to reach customers, resulting in higher engagement and repayment rates.
This level of smarter profiling and targeting of borrowers increase not only the likelihood of reaching the borrower but also the recovery rates for each targeted individual.
P2P lending platforms can stand to benefit from tech-driven debt collection, but the platform can also find a home within a variety of organizations, from collection agencies and digital lenders to banks and non-bank institutions. A more human-centered and targeted approach to the way we recover debt also reduces a company's exposure to any form of reputational risk.
Perhaps as an industry, we needed to go through these terrible growing pains to truly understand the double-edged sword we have allowed into the market. To answer the question posited above though – yes, I do believe that these tragedies could have been avoided.
Industry players need to envision a holistic application of P2P lending into new markets, with all stages of a borrower’s life cycle brought into consideration.
Front and center of these efforts is one crucial question: do we truly understand the underserved markets that require P2P lending?
I think that once we do, the rest will follow naturally.
Guillermo Martin is the head of global sales and Indonesia country manager at Asia Collect, a Singapore-based fintech company that aims to reform the collections industry using AI and machine learning.