Debt Collection goes Private

Jul 18, 2019

With the IRS claiming between $50-$52 billion in collectable tax debt, but limited resources to collect it, Congress passed a law requiring the IRS to use private collection agencies to help retrieve this debt. The law, passed in 2015, led the IRS to assign certain types of inactive debt cases to Private Debt Collection (PDC) agencies, in order to help the IRS collect the debts they were not actively pursuing. An analysis by the GAO (Government Accountability Office) has deemed these Private Debt Collection programs somewhat ineffective when it comes to return on investment. After a lengthy analysis of the IRS’s PDC program, by the GAO, evaluating the program for safeguards to protect taxpayers from risks such as scammers impersonating collection agencies, gaps in the current process were noted.Those gaps included the need for the IRS to improve PDC program objectives and measures, revenue and cost reporting, analysis to assign cases, and management of taxpayer risks.

Over 700,000 debt collection cases were assigned to private collection groups in 2018, collecting nearly $89 million dollars with a cost of $67 million to the IRS. This PDC program saw financial success, though $51 million of the collected revenue went to The Treasury and the remaining $38 million was put toward current and future program funding, leaving Congress with an incomplete picture of the program’s true costs and return on investment. The GAO estimates that approximately 34% of collection cases result in recovered funds--citing not being able to contact the taxpayer and inability to collect the debt as barriers in collecting funds. Utilizing advanced analytic programs, the PDC program could more effectively identify cases that would result in collection. Some of the most significant advances brought about by advanced analytics and machine learning are in customer segmentation, which is becoming much more sophisticated and productive. Employing machine learning algorithms that collect data on successful collection attempts and continue to evaluate the data to determine the best collection attempt cases, can allow PDC agencies to move to a deeper, more nuanced understanding of their at-risk customers. Segmenting customers into effective treatment streams can provide collections agents with more focused, streamlined, targeted cases allowing agents to utilize interventions more likely to prove successful for customers in those segments.

Though the IRS is working to improve safeguards to help avoid risks to the taxpayers from PDC programs, the GAO deduced there were additional risks the IRS had not considered (including taxpayers signing up for payment plans they could not afford). It was reported that more than 10,000 taxpayers had been scammed out of roughly $55 million between 2013-2017 by people pretending to be IRS collection enforcement. The IRS has targeted six specific taxpayer risks related to the PDC program, including these scam calls, and is working to publicize how to avoid scams. In addition, though the IRS monitors collection agency calls and complaints, it does not analyze whether the response to risks are effective. Machine learning and nontraditional data have become the new frontier in collections-decision support. By allowing algorithms to work through thousands of audio conversations, PDC agencies can discover the most productive approaches. With suggested approaches, informed by insights from the field of behavioral science, machine learning can be used to diagnose and neutralize the biases that affect collector and customer decision making. At the same time, the machine-learning approach is enabling automation of larger classes of decisions. The main task of the PDC program is to close the tax gap and promote tax compliance, all while maximizing resource use and protecting taxpayers.

ASR’s RevHub Collections Accelerator uses advanced analytics, like machine learning, to segment accounts into workstreams that allow agencies to utilize their workforce and resources most effectively. Click here to learn more about RevHub.



Tags: