CWRTP Data Analysis Leads to Increased Arrears Collections

What if the State of Iowa could help reduce the number of child support cases in arrears? What if there was a way to predict which individuals are more likely to pay off arrears in the future? These two questions guided a joint, two-year data mining project between the Iowa State University (ISU) Child Welfare Research and Training Project (CWRTP) and the Iowa Department of Human Services (DHS) Child Support Recovery Unit (CSRU) and led to the collection of more than $314,000 in IV-D payments during a three-month period in 2016.

Child support arrears are past due child support payments that a non custodial parent—or the parent without primary custody of the child—owes to the custodial parent. The purpose of the ISU-CSRU arrears project was twofold: (1) to design a new system or tool with better indicators to help reduce existing arrears and prevent the accumulation of future arrears; and (2) for ISU to assist CSRU in prioritizing arrears cases based on the non custodial parent’s likelihood to pay.

Most of the work was completed on the ISU campus by staff and graduate students in close collaboration with CSRU staff. Meetings were conducted through email, conference calls, and face-to-face consultations. ISU staff and students played a key role in the literature review, data analysis, and creation of statistical modeling procedures, while CSRU staff provided several sets of data for the project, including information on arrears payments from fiscal years 2010-12 and 2014-15.

Feng Zhao

Feng Zhao is a doctoral student in the Department of Human Development and Family Studies and was the primary researcher on this project.

Working with CWRTP Director Dr. Janet Melby, Feng Zhao (pictured), a doctoral student in the Department of Human Development and Family Studies, performed statistical analysis on the datasets, excluding data related to interstate cases and cases with incarcerated payors. After identifying key variables in both paying and non-paid arrears cases, Zhao built two analytical models: (1) one that could predict whether a case got paid in fiscal year 2013, which, in turn, allowed him to determine the probability of cases receiving payments (PGP); and (2) another predicting the variables associated with higher payment in fiscal year 2013, which then allowed him to obtain a regression equation that could be used to calculate the expected annual payment (EAP) for non-paid cases. The final step was to create a “priority table” using both PGP and EAP to categorize arrears cases according to likelihood of payment, as well as the amount of expected payment.

As a result of this analysis, CSRU was able to assign two “clean-up” projects to field workers that ultimately allowed the organization to collect payments on high-priority arrears cases. In September 2015, the first clean-up project asked workers to review 264 non-paying cases where it appeared the payor was receiving social security disability (SSD) or social security retirement (SSR) benefits. After reviewing the 264 initial cases, it was found that 221 cases had payors who were still receiving SSD or SSR benefits.  After further review of the 221 cases

  • Workers sent income withholding notices on 30.77% of the cases,
  • As a result of sending these income withholding notices, 29.86% of the cases received an income withholding payment.

The second clean-up project was intended to determine whether the verified employer listed was still accurate on the non-paying arrears cases.  This project  was sent to the field on April 22, 2016. As of August 1, 2016, the 3,294 active cases showed:

  • The workers removed the employer on 52.70% of the cases as they were no longer valid.
  • On the remaining cases where the employer was valid the workers called and/or sent the employer a new income withholding order. Due to this effort, 612 cases received a payment.
  • From April 22 to July 20, 2016, these cases have collected a total of $314,460.58 in IV-D payments

“After completing these clean-up projects, we discovered that the cases where clean-up was done were the same cases with the highest priority scores and had already been worked,  Many of these cases were now receiving payments as a result of the hard work of our field staff,” explained CSRU program planner Shannon Thill. More importantly, though, she notes that “through this project, we were able to locate sets of cases that could be worked easily and updated by field workers in order to increase our arrears collections.”

Ultimately, the project succeeded thanks to extensive collaboration between CSRU staff with their knowledge of field operations and CWRTP data analysts—graduate assistants Feng Zhao, Dong Zhang, and Chen Peng, Research Analyst Erkuan Wang, and Director Jan Melby—who developed the predictive models.

The arrears project was conducted as part of the contractual partnership between Iowa State University and the Iowa Department of Human Services. Providing oversight for the partnership are Dr. Gong-Soog Hong, Principal Investigator at ISU, and Carol Eaton, Bureau Chief at the Iowa DHS Child Support Recovery Unit.