The Pensions Regulator in the U.K. has teamed up with the Government Digital Service to help it foster good governance by retirement plans in terms of protecting their participants.
The pension plan watchdog will use a machine learning model, built by GDS data scientists, to predict behavior by retirement plans. Using decision-tree algorithms, which have a flow-chart structure, scientists will track details of retirement plan behavior such as how many times officials have contacted TPR in the past year, and whether a pension fund's investment returns were submitted to TPR on time.
TPR will split the data into about 30 different groups based on retirement plan characteristics, including how they report their returns, according to a blog by Peter Jackson, TPR's former head of data. Mr. Jackson left the regulator in April, but TPR is committed to the project, a TPR spokesman said.
This article originally appeared in the May 15, 2017 print issue as, "U.K. watchdog uses algorithms to track funds' governance".